Quranic Studies; Sources And Methods Of Scriptural Interpretation Wansbrough

John Wansbrough

Quranic Studies

Sources and Methods of
Scriptural Interpretation

Quranic Studies

Many believe that the academic study ol the
Qur'an lags far behind the study ol the
Bible while being, at the same time, closely
modeled alter it. Not only are the relevant
scholarly resources of the Qur'an less nu-
merous than those available in biblical
scholarship, but comparatively speaking
the variety of methods employed to deal
with the scriptural text has been severely
limited. One of the first groundbreaking
efforts in Islamic studies was made by John
Wansbrough in his unique work Quranic
Studied: Source.' and Method' oj Scriptural In-
terpretation. Written between 1968 and 1972,
this revolutionary analysis had a profound
effect on the study of Islam. It produced,
in the minds of many, a wholly new dichot-
omy in the approach used in Islamic stud-
ies: on one side, the skeptical revisionists
and on the other, the trusting traditional-
ists. Well ahead of his time, Wansbrough
questions the very basic assumptions ol
previous scholars in a way that had never
before been attempted. Working with the
heritage of Joseph Schacht and Ignax Gold-
ziher before him, Wansbrough approached
the Qur'an in a manner that sees the Mus-
lim tradition as grounded in the dogmas ol
later centuries. Freed from these con-
straints, new questions relevant to contem-
porary scholars had to be asked.

Wansbrough was the first to analyze
carefully the documents from the first four
centuries of Islam that describe the rise
of the Qur'an to the position ol absolute au-
thority in the Muslim community. Al-
though these works were known to exist,
no modern scholar had actually read them
and tried to make coherent sense of the
material. Wansbrough carved out new areas
of inquiry and debate for scholars and lay
enthusiasts alike.

continued on /><«■/■ //<i/>

John Wansbrough

Quranic Studies

Sources and Methods of
Scriptural Interpretation

Foreword, Translations, and Expanded Notes by

Andrew Rippin

Prometheus Books

59 John Glenn Drive
Amherst, New York 14228-2197

Published 2004 by Prometheus Books

Quranic Stndi&i: Sourcej and Methodj of Scriptural Interpretation. Copyright © 2004 by
Maria Emma Theresa Wansbrough. All rights reserved. No part of this publication
may be reproduced, stored in a retrieval system, or transmitted in any form or by any
means, digital, electronic, mechanical, photocopying, recording, or otherwise, or con-
veyed via the Internet or a Web site without prior written permission of the publisher,
except in the case of brief quotations embodied in critical articles and reviews.

Inquiries should be addressed to
Prometheus Books
59 John Glenn Drive
Amherst, New York 14228-2197
VOICE: 716-691-0133, ext. 207
FAX: 716-564-2711
WWW.PROMETHEUSBOOKS.COM

08 07 06 05 04 5 4 3 2 1

Library of Congress Cataloging-in-Publication Data

Wansbrough, John E., 1928-2002

Quranic studies : sources and methods of scriptural interpretation / John
Wansbrough ; foreword, translations, and expanded notes by Andrew Rippin.
p. cm.

Originally published: Oxford : Oxford University Press, 1977, in series: London
oriental series.

Includes bibliographical references and index.
ISBN 1-59102-201-0 (alk. paper)

1. KorarT— Criticism, interpretation, etc. — History. 2. Arabic language —
History. I. Title.

BP130.45.W36 2004
297.1'226-dc22

2004040088

Printed in the United States of America on acid-free paper

CONTENTS

FOREWORD

pageix

Andrew Rippin

PREFACE

xxi

ABBREVIATI ONS

XXV

BIBLIOGRAPHY

xxvii

MANUSCRIPTS UTILIZED IN QURANIC STUDIES

xxxix

Andrew- Rippin

REVELATION AND CANON

1

1. lne document

1

2. Its composition

33

FMRT FM9 OF PRnPHFTHODD

ORIGINS OF CLASSICAL ARABIC

85

rKlNL.lrLto vJr fcXtAjhMo

119

i. Haggadic exegesis

122

2. Deutungsbediirfiigkeit

I4O

3. Halakhic exegesis

170

4. Masoretic exegesis

202

5. Rhetoric and allegory

227

INDEX

247

A. Names and subjects

247

z 4v

C. Quranic references

252

ANNOTATIONS

257

Andrew Rippin

GLOSSARY

309

Andrew Rippin

FOR MY WIFE

Quranic Studies

FOREWORD

THE academic study of the Qur'an, it has often been remarked, lags far
behind the study of the Bible while being, at the same time, closely
modelled after it. Not only are the resources available to scholars of the
Qur'an much more limited than those available to their biblical-scholar
counterparts, but the depth of methodological experimentation in
dealing with the scriptural text has been severely limited in compari-
son. This situation is illustrated by consideration of the sheer quantity
of scholarship that has been produced and the number of scholarly
landmarks that exist in the field. Modern biblical scholarship fills a
library many times the size of that devoted to the Qur'an. Each subdis-
cipline of biblical studies has its own set of "classics." By contrast, it is
still possible to point to individual works in the history of the study of
the Qur'an and declare them the pivotal texts that provide the founda-
tions for all later studies.

Two nineteenth-century authors set the tone, perspective, assump-
tions, and approach for much of modern scholarship on the Qur'an.
Abraham Geiger (d. 1874) is most famous as the founder and leader of
the German Jewish Reform movement. In 1832 Geiger submitted a
contest entry (written in Latin) to the University of Bonn that was
published the following year in German under the title Was hat Mo-
hammed aus dem Judenthume aufgenommen? 1 Geiger's approach to
the Qur'an marks the beginning of the European scholarly quest for
the sources of Muslim scripture in Judaism and, to a lesser extent,
Christianity. No longer was the Qur'an being approached from the
medieval perspective of polemic grounded in the notion that
Muhammad was a religious impostor. Geiger's work set a new direc-
tion for scholarship because its working assumption was that
Muhammad was sincere in his religious mission. Geiger's study was
motivated by the underlying thrust of post-Enlightenment work gener-
ally, which promoted a sense of curiosity to which no particular value
was added over and above the desire to know the previously unknown.

Solid philological scholarship also serves as the foundation for in-
vestigation of the Qur'an, and the work of one of the greatest

1 Translated into English by F. M. Young under the title Judaism and Islam (Madras,
1898; reprint, New York, 1970).

X

FOREWORD

philologists of the Semitic traditions, Theodor Noldeke (d. 1930),
provides the second major point of departure for scholarly studies of
the Qur'an. Noldeke's lasting monument to Qur'anic scholarship, his
Geschichte des Qorans, has a history similar to that of Geiger's work.
Written originally in Latin, it was submitted in 1856 as a dissertation
and awarded the winning prize in a Parisian competition for a study of
the "critical history of the text of the Qur'an." The work was first
published in an expanded German edition in i860. A second edition
of the work appeared in three volumes, with volumes 1 and 2 edited
and rewritten by Friedrich Schwally (1909, 1919) and volume 3 writ-
ten by Gotthelf Bergstrasser and Otto Pretzl (1938). Noldeke's work
has set the agenda for subsequent generations of Quranic scholarship
by emphasizing concerns with chronology in the text and the text's
biblical background. As well, Noldeke's philological insights provide
much of the lasting value; his treatment of language, his stress upon
etymology, and his insights into grammar all provide the model for the
philological study of the Qur'an, and the material he provided contin-
ues to be a valued source of reference for later scholarship.

Such are the scholarly foundations of the field of study. However,
several works appeared in the 1970s and 1980s that deeply affected
scholarly approaches and attitudes toward the study of the Qur'an in
the contemporary period. Each work, in its own way, opened up a new
range of approaches, a new inventory of questions to be asked, and a
new selection of paths to be followed by the following generation of
students of the Qur'an. One major development is seen in the work of
Toshihiko Izutsu, whose books have been praised by Muslim and non-
Muslim scholars alike. While a parallel methodological movement was
taking place at the time in biblical studies, 1 Izutsu's works, The Struc-
ture of Ethical Terms in the Koran (Tokyo, 1959) and God and Man
in the Koran (Tokyo, 1964), as well as the rewritten version of the
former work, Ethico-religious Concepts in the Qur'an (Montreal,
1966), appear uninfluenced by research in the biblical field (in either a
theoretical manner or in matters of detail). 2 The study of the Qur'an
was, with Izutsu, moved out of its biblical context and situated in an
explicit methodological framework of semantic analysis. Set out by

1 Represented most significantly in James Barr, Semantics of Biblical Language (Oxford,
1961)-

2 The chronological development of the two fields at roughly the same time is perhaps
best understood in light of developments in the field of linguistics. See Stephen UHmann, The

Principles of Semantics, 2nd ed. (Oxford, 1957).

FOREWORD

XI

Izutsu at the beginning of his Ethico-religious Concepts in exemplary
form, the method is based upon one core idea: that the meaning of
words inheres in the context of their use, not in the individual words
themselves. Thus, on one level, this approach is in direct opposition to
the philological method, given the latter's stress on etymology in
determining meaning. While other assumptions about the Qur'an are
shared by the semantic approach with earlier approaches (continuing
the presuppositions of the Muslim tradition about the history of the
text, for example), the influence of the semantic method is apparent in
the many studies of individual words and word patterning that have
followed in the wake of Izutsu 's work.

Still, there are major questions that linger in the study of the
Qur'an, especially related to why the text looks the way it does. Recent
works dealing with these problems have provided increasingly system-
atic attempts to devise a solution/ Undoubtedly the most successful
and influential initiative is seen in the work of Angelika Neuwirth.
Neuwirth's book, Studien zur Komposition der mekkanischen Suren
(Berlin, 1981) argues for formulaic construction techniques structuring
the Qur'an that reflect a liturgical, oral background of "cola" as build-
ing blocks for the text. 2 That observation leads to arguments in favor of
the Qur'an being seen as a carefully composed literary structure. Neu-
wirth's later work has expanded upon that insight in emphasizing the
liturgical and heavily symbolic nature of the language of the Qur'an.

A third work that has significantly affected the field of study is the
one being reprinted here, John Wansbrough's Quranic Studies: Sources
and Methods of Scriptural Interpretation. Secondary scholarly com-
mentary on this book is readily available, 3 and the following remarks

1 Earlier attempts may be noted: D. H. Miiller, Die Propheten in ihrer urspriinghchen
Poesie (Vienna, 1896); R. Geyer, "Zur Strophik des Qur3n," WZKM2.2. (1908). 265-86.

2 "Cola" are also discussed in A. H. Mathias Zahniser, "The Word of God and the Apos-
tleship of 'Isa: A Narrative Analysis of Al 'Imran (3):33— 62," 75536(1991): 77—112; "cola" are
defined as "breath units of speech, the number of syllables which can be encompassed in a
single breath" (p. 88).

3 A. Rippin, "Literary Analysis of Qur'an, Sira and Tafsir: The Methodologies of John
Wansbrough," in Approaches to Islam in Religious Studies, ed. R. C. Martin (Tucson, AZ,
1985), pp. 151—63, 227-32 (reprint, Oxford, 2001); volume in honor of J. E. Wansbrough,
BSOAS 57, no. 1 (1994), including a full bibliography of Wansbrough's writings, pp. 4—13;
Herbert Berg, ed., "Islamic Origins Reconsidered: John Wansbrough and the Study of Early
Islam," special issue, Method and Theory in the Study of Religion 9, no. 1 (1997); Herbert
Berg, "Competing Paradigms in the Study of Islamic Origins: Qur'an 15:89-91 and the Value
of isnads," in Method and Theory in the Study of Islamic Origins, ed. H. Berg (Leiden, 2003),
pp. 259-90.

xii

FOREWORD

highlight just a few of issues to help the reader appreciate the signifi-
cance and uniqueness of the work. Written between 1968 and 1972,
this book has had an impact that has been felt throughout many differ-
ent areas in the scholarly study of Islam. It produced, in some people's
minds, a dichotomy in approaches among scholars, between the skepti-
cal revisionists and the trusting traditionalists, that has influenced the
entire field of Islamic studies. Such a bifurcation, though convenient
for polemical purposes, hardly corresponds to the methodological
diversity and independence of those who work in the area. Fundamen-
tally, Wansbrough's work questions the basic assumptions of previous
scholarship in a way that had not been attempted before; working with
the heritage of Joseph Schacht 3 and Ignaz Goldziher 4 before him,
whose work had focused on issues of Muslim law, Wansbrough ap-
proached the Qur'an in a manner that sees the Muslim tradition
surrounding the text as grounded in the dogmas of later centuries.
Freed from those constraints, new questions can, and must, be asked.
Wansbrough's work was also stimulated by advances in research in
manuscript studies, especially the publication of Fuat Sezgin's
Geschichte des arabischen Schrifttums, vol. 1: Qur'anwissenschaft —
Hadlt — Geschichte — Fiqh — Dogmatik — Mystik bis ca. 430 H. (Lei-
den, 1967). Sezgin's approach to the early history of written
documents in Islam (as also illustrated in his own earlier work on
Bukharl) is reflected in his cataloging of the manuscript treasures that
had not previously been easily accessible to scholars; however, ascrip-
tion of works to early Muslim figures (such as Ibn 'Abbas, d. 68
AH/ 687 CE) tends to be accepted by Sezgin and then used as evidence
of the written transmission of documents in the first few Islamic centu-
ries.

Wansbrough was the first person to subject to scholarly analysis an
entire body of literature attributed to the first four centuries of Islam
that stands as a witness to the rise of the Qur'an to the position of
absolute authority in the Muslim community. Although these exegeti-

1 This is according to a letter I received from Wansbrough, June 12, 1980; he also re-
ported that Sectarian Milieu was written between 1973 and 1977.

2 For example, see J. Koren and Y. D. Nevo, "Methodological Approaches to Islamic
Studies," Der Islam 68 (1991): 87—107. For an excellent overview of the methological debate,
see Herbert Berg, The Development of Exegesis in Early Islam: The Authenticity of Muslim
Literature from the Formative Period (Richmond, UK, 2000).

3 Especially Joseph Schacht, The Origins of Muhammadan Jurisprudence (Oxford, 1953).

4 Especially Ignaz Goldziher, Muhammedanische Studien (Halle, 1889—90); translated as
Muslim Studies, 2 vols. (London, 196,6, 1971).

FOREWORD

Xlll

cal works were known to exist, having faithfully been cataloged by
Sezgin, no scholar had actually read them and tried to make coherent
sense of the material. This is one of Wansbrough's main accomplish-
ments as reflected in this book, which lists seventeen manuscript
works. Notably, almost all of these books have now been edited and
published. Although Wansbrough may have provided some of the
impetus toward this publishing industry, the general rise in the publi-
cation of works of Arab heritage is far greater and more a part of an
overall social tendency in the modern Muslim world than what can be
accounted for by his sheer influence on this particular matter.

Wansbrough's early scholarly career focussed on Judaeo-Arabic and
Mediterranean trade history. His move into Qur'anic studies resulted
from his interest in literature in general and especially the literary traits
of documents from the medieval times. Given Wansbrough's path into
the study of the Qur'an through an emphasis on the cultural produc-
tion of literature, it should not come as a surprise to any of his readers
that he enjoyed the use of the English language and its rich classical
heritage as much as he did that of Arabic. His enjoyment of language is
likely the cause of one of the most frequently voiced complaints about
this work — that it is "difficult." In reflecting upon this matter, I must
intrude a personal anecdote. For reasons I no longer remember, I had
occasion, in my eagerness as a graduate student, to talk to Wansbrough
about Malcolm Lowry and Under the Volcano. Early on he may well
have recommended that I read it. For my own part, I was able to show
him some photographs that I had taken of the memorial to Lowry in
Dollarton in North Shore Vancouver, on the site at which Lowry wrote
the book. Our conversation turned to Wansbrough's own work, and I
remember mentioning to him that people found Quranic Studies
"difficult." He looked at me with an expression which suggested both
puzzlement and exasperation. "What's the problem with being diffi-
cult? Lowry is difficult, isn't he?" It is also worth remarking, to
emphasize the point further, that Wansbrough wrote fiction as well, at
least one piece of which was published, in 1980.

1 See the section below, "Manuscripts Utilized in Quranic Studies," for details. Farra',
Ma'am 'I-Qur'an, is listed in the bibliography in both its published and manuscript versions;
Wansbrough appears to have mainly used the printed edition.

2 John Wansbrough, "Let Not the Lord Speak," Encounter 54, no. 5 (1980): 3-7. In his
letter to me of June 18, 1980, Wansbrough also reports that he had written a novel entitled
"Palimpsest" in the period March— September 1979.

xiv

FOREWORD

There is no doubt that Quranic Studies needs serious, attentive
study, just like any other stimulating work of scholarship. It is not a
light read. It is also allusive in some of its master narrative statements
to the biblical and Arabian backgrounds, a fact that has captured many
people's attention. It is those points that have led to many exchanges
and debates among scholars and others, most of them unprofitable in
my view, regarding the most "outrageous" claim that the Qur'an was
written (down) in the third-century hijri. Indeed, some people seem to
have reacted to Quranic Studies solely on the basis of this matter, as
though Wansbrough wrote the book in order to prove that point but
that he buried it within the text in an obscure manner in order to
lessen the impact. To me, this is a misreading of the book: the subtitle
of the work, Sources and Methods of Scriptural Interpretation, summa-
rizes perfectly the aim of the work. That a reading of the
interpretational tradition suggests that the Qur'an emerged late is
simply a result of that investigation, but it is only one of many. Much
of the reaction to this suggestion of Wansbrough illustrates that some
scholars continue not to question their own scholarly assumptions.
However, the critical issue here that Wansbrough's work provokes in
me as a reader, at least, is, what do we mean by "the Qur'an" when we
speak of it in this way? A scripture has many different aspects to it
beyond its apparent written form, and even the written form takes time
to evolve and stabilize (as Muslim tradition has always maintained).
Drawing attention to this question is one of the major merits of Wans-
brough's work.

It is worth pondering this critical issue of the emergence of the
Qur'an at greater length. Fundamentally, the question revolves around
what we mean by "the Qur'an" and what sort of evidence we have to
answer that question. There is a good deal of uncertainty as to what we
do mean here by "the Qur'an." Wansbrough is certainly talking about
something that has significance, not a theoretical construct. I would say
that when we speak of the Qur'an in this context, and if we are going
to have a meaningful discussion of the question, three elements must
come into play: one, there must be a fixed body of text that is, two,
written down, and, three, has some measure of acceptance among a
group of people as a source of authority.

1 I previously presented some of what follows at the symposium on "Who Wrote (Down)
the Qur'an?" at the University of Philadelphia, February 2003.

FOREWORD

XV

Now, those processes of being fixed, of being written down, and of
becoming authoritative, work hand-in-hand. Significantly, it is precisely
these three notions — proliferating copies of the Qur'an, with a com-
monality of content, being promulgated as having symbolic and
practical authority — that are, of course, the central motifs of the com-
munity's own account related to the collection of the Qur'an under
Abu Baler, TJmar, 'Uthman and 'All. Notably, those are precisely the
key elements that are central to the criteria of what we need to talk
about for a meaningful discussion of this text called the Qur'an. On the
level of characteristic motifs, then, I doubt that there is much dis-
agreement.

There are certainly tensions within the community's account of the
collection of the Qur'an when it is subjected to scholarly scrutiny,
especially in the matter of the speed at which this process of definition
of the limits of the text is conceived to have occurred. This is even
more so when one considers the significant social processes that must
have gone on behind each of the elements in this chain of events. For
example, the emergence of a fixed and explicit script for the Arabic
language suggests a considerably longer time frame than the commu-
nity's account, on its surface, seems to allow. But it is also worthy of
note that these internal community accounts of the collection of the
Qur'an do not, in fact, fully account for the establishment of the
Qur'an even within the community's own view. Elements such as the
fixing of the Arabic script, the establishment of written formats of
manuscripts, and the codification of the variant readings are all spoken
of in these accounts as subsequent to the initial collection stories by a
significant time period. Since each of these elements is important in
establishing the authoritative, fixed, widely distributed text of what we
know as the Qur'an, this is of substantial significance. The communal
process of text formation is a long one, and a notion of the Qur'an
"existing" under 'Uthman, for example, is quite anachronistic, if we
give full meaning to the notion of the "Qur'an."

Furthermore, it is also worthy of note that the evidence of the ex-
isting Qur'an manuscripts does not match the rationale provided by the
community's collection stories. The creation of a written text in order
to solve the problem of discrepancies in the reading of the text, as is
suggested as the rationale for the distribution of copies of the text
under 'Uthman, is clearly an anachronistic element. The form of the
Arabic script in early manuscripts is far too ambiguous to have accom-
plished the purpose of unifying the reading of the text.

xvi FOREWORD

Now, it may be thought that I have defined the question of what
the Qur'an is in an unreasonable way and that I am speaking, in fact,
of something that all agree is a later product. However, I think that is
not the case. Those who consider the Qur'an to have been written
down some thirty or fifty years after the death of Muhammad and who
thus speak of a Qur'an existing under 'Uthman frequently wish to
ascribe or presume both an unquestioned stability and an absolute
authority to the text at that point. Such stability and authority seem to
me to be unrealistic, in tension with other sources that provide us with
glimpses into the process by which the Qur'an came into being, and
substantially contradicted by the state of the text at that point of his-
tory. To speak, using Fred Donner's words, of the time "when the
whole Qur'an as a canonized closed text first crystallized" being so
early in the community's existence does not fit with all of the evidence
that is available to us. It is this questioning of previously unchallenged
scholarly assumptions that is Wansbrough's major achievement.

Wansbrough pushes these questions further. He argues in Quranic
Studies that the commentarial literature provides us with material that
can help answer questions such as how a fixed text of the Qur'an was
made acceptable within the community and how it was defined in an
age that depended upon handwritten copies. These matters are posed
within the context of the process that we commonly call canonization.
Such a process differentiates between the composition of a text and its
recognition as scripture, with all the implications of that term. In fact,
the process is sometimes spoken of as having five stages: composition,
circulation, revision, collection, and recognition. Many factors play into
that final "recognition" factor, but the opinion of religious leaders and
the conviction of common people clearly play a major role. This stage
of recognition cannot be pinned down; it does not occur overnight as
the result of one person's actions or even the actions of a group of
people. Canon formation is a part of a process of knowledge dissemina-
tion and control within an institutional setting. In the Islamic case, the
actions of the scholarly elite — the jurists, the grammarians, the theolo-
gians — whose actions work to define the texts that are to be studied,
reflected upon, and promulgated are crucial, but, once again, canoniza-
tion is not a single process, nor is it a personalized one. One of the
marks of this canonization process, according to Wansbrough, is the

1 Fred M. Donner, Narratives of Islamic Origins: The Beginnings of Islamic Historical
Writing (Princeton, NJ, 1998), p. 37...

FOREWORD

xvii

emergence of commentarial literature: the subjecting of a text to a
process of elaboration and definition within the understanding of the
community. It is this process that creates and confirms the authority of
the text.

This point is crucial. The fundamental argument is that we cannot
meaningfully talk about the Qur'an as we know it today until that
point of authority, acceptance, and stability has been achieved. That
does not mean, of course, that a proto-Qur'an did not exist, in some
form or another, prior to that point. Indeed, the text must have a
prehistory for such a process to take place, a prehistory that brings
strands of the earlier biblical and Arabian traditions together through
the person of Muhammad. However, to return to the central point
here, the evidence that these commentaries on the Qur'an provide is
one element in the overall puzzle of the emergence of the scripture.
The organization, solidification, and standardization of the text of the
Qur'an — crucial elements in creating what we know as the Qur'an —
are accomplished, confirmed, and emphasized in this series of early
commentarial texts. Of course, such historical observations must still fit
within the context of all the other evidence that can be brought forth
to deal with the core question of the emergence of the Qur'an. Scholars
subsequent to Wansbrough have brought forth all sorts of further
evidence arguing the point on one side or the other. The simple exis-
tence of written portions of the Qur'an in manuscript form, for
example, does not tell us the full picture of the history of the text any
more than the traditional collection stories or the language and struc-
ture of the text itself do.

Fred Donner has written in his recent book that we must be able
"to say who was responsible for deciding what did, or did not, belong
to the Quranic canon." He continues, "To put the responsibility for
such a process simply on 'the community' or 'the scholars' is too
vague; we need to have some idea of what individuals, or at least what
groups, were involved in making decisions, and what interests they
represented." However, at this point in the history of the critical
scholarly study of the Qur'an, our knowledge does not really permit us
to move beyond these admittedly vague notions; the facts remain
suggestive of a multitude of ideas to us, just as they do in Wans-
brough's work. We do not know enough about the social, polemical,
and political processes taking place during the first centuries of Islamic

1 Ibid., pp. 37-38.

xviii FOREWORD

civilization; we have only the vaguest information about the develop-
ment of professional classes, the religious scholars, and the social
power structures. The current state of our knowledge simply does not
allow us to specify the who, when, or where. We need much more
investigation into the nature of early Islamic society before it will be
possible to say more. The failure to be able to answer those questions
cannot be taken as evidence against the relevance of the implications of
the various sets of evidence that can be brought forth as they are in
Wansbrough's work; rather, such a vague answer to the question
indicates by its very nature the amorphous and collective sense within
which all such canonizing processes take place.

With even such a tentative conclusion in place, however, we can
start asking subsidiary questions related to the Qur'an, ones that are
often confused with ideas related to the emergence of the authoritative
text. Clearly, such a picture of the process lying behind the emergence
of the crystallized, canonized text of the Qur'an puts us in a position to
ask questions about the compositional history of the Qur'an. Here, too,
Wansbrough makes us rethink the assumptions regarding this. Is the
Qur'an to be attached fully to Muhammad? Is there evidence that the
Qur'an was reformulated, added to, changed, or developed during that
process of canonization? The first parts of Wansbrough's book address
these points, but they, too, lead to the final crucial questions sketched
above: What was the process behind deciding the contents of the
Qur'an? How were the limits of the text established (such as to exclude
the hadith qudsi, for example)? How were those processes confirmed
and conveyed? What was the process by which the meaning of the text
was defined within certain limits, especially as connected to the matter
of variant readings?

Some people have decried the publication of Wansbrough's
Quranic Studies, seeing it as a major impediment to fostering a trust of
nonconfessional scholarship among Muslims. Such an accusation made
less sense when the work was published than it does, in fact, today. A
book with a small print run, written in a language whose tone and
vocabulary was scholarly through and through, the audience for
Quranic Studies was small. Today, the Internet has changed that
substantially. Used as a source of authority and opinion on polemical

1 See the work of Norman Calder, Studies in Early Muslim Jurisprudence (Oxford,
1993), esp. chapter 7, "Literary Form and Social Context," for an illustration of what a study
on this topic might entail.

FOREWORD

XIX

sites by both Muslims and Christians, Wansbrough's work has gained
a significant profile, even though likely languishing unread by those
citing the work. This is, to me, a most unfortunate fate for such a
stimulating work of scholarship, and I believe Wansbrough himself
would have bemoaned the fact. Given the situation, however, I also
believe that the best way to counter such polemic and nonscholarly
treatments of Wansbrough's ideas, extracted, as they are, from their
context and manipulated in various ways, is to have the book become
more widely available for consultation: hence this reprint. Honest
scholarship (of which I certainly believe Wansbrough's work to repre-
sent the highest level) has nothing to fear from open discussion and
debate among those who are informed.

Andrew Rippin
Victoria, British Columbia
December 2003

r

PREFACE

Despite long reflection, many false starts, and unceasing efforts at clearer
formulation, these studies remain essays. Time and industry, together with
the mechanism of cross-reference, have served perhaps to make of them
one essay, but not to eliminate the basic etymological sense of that term.
This estimate, intended neither to disarm nor to discourage, is the conse-
quence as much of an experimental method as of an inordinate quantity of
literary material to be examined. To argue a case for the Qur'an as scripture
may seem gratuitous. As the record of Muslim revelation the book requires
no introduction. As a document susceptible of analysis by the instruments
and techniques of Biblical criticism it is virtually unknown. The doctrinal
obstacles that have traditionally impeded such investigation are, on the
other hand, very well known. Not merely dogmas such as those defining
scripture as the uncreated Word of God and acknowledging its formal and
substantive inimitability, but also the entire corpus of Islamic historio-
graphy, by providing a more or less coherent and plausible report of the
circumstances of the Quranic revelation, have discouraged examination of
the document as representative of a traditional literary type. But historio-
graphy, like other kinds of literature, derives an important share of its
momentum from the rhetorical devices upon which it depends for ex-
pression, that is, upon techniques designed, developed, or borrowed to
enhance and to interpret its communication. Historical reports of the
Quranic revelation are no exception, and it seemed to me that a structural
analysis, not only of the text of scripture but also of the other evidence
associated with its genesis and with its interpretation, might produce some
useful comparisons with the traditional historiography.

I have thus proposed for the study of Islamic monotheism a threefold
division into its principal components: scriptural canon, prophetology, and
sacred language. Examination of these, in the first three chapters of this
book, will, I hope, provide an adequate introduction to my main concern,
which is the development of scriptural exegesis. There the treatment is
typological, though I have sought to elicit an at least provisional chronology.
Provisional too are my views on the elaboration of the canon, on its precise
relation to the much larger corpus of prophetical literature, and on the
environment which produced the final forms of both. Such problems as
attend a typological description of the traditional components of salvation
history of the decisive factors in community formation, the boun-
daries between orthodoxy and heterodoxy resulting from formulation of a
fixed identity and elaboration of a historical image, and the articulation

xxii

PREFACE

of a dogmatic theology as final stage and permanent emblem of the reli-
gious establishment, await clarification and eventual (tentative) solutions.
These may be found, I suspect, in the Islamic literature of polemic, both
sectarian and interconfessional, and it is my intention to devote a second
volume to analysis of that literature. The interesting hypotheses of such
scholars as Schoeps and Rabin, following upon the theories of Harnack
and Schlatter, with regard to the sectarian catenations from which Islam
in the end emerged, or constituted the final expression, deserve notice.
There the essential difficulty lies in identifying the several shared theo-
logoumena as products of historical diffusion rather than of polygenesis.
All such efforts at historical reconstruction (zoie es eigentlich gewesen) tend to
be reductive, and here one senses the spectre of that (possibly very real)
dichotomy in early Christian history: Jerusalem Urgemeinde opposed to
Hellenistic kerygma (Bultmann). The basic problems associated with that
opposition, whether social or doctrinal, seem in retrospect to reflect
disputes about eschatology , much as the development of Rabbinic Judaism
has been defined as reaction to or residue from extreme expressions of
eschatological belief/activity. Now, in the formation of the Muslim com-
munity an eschatological factor is hardly perceptible, though much has
been made of the presence in Muslim scripture of eschatological formulae
(Casanova and Andrae). In my view it is unlikely to have been an eschato-
logical community which served either as model or as point of departure
for Islam, but rather one of or a combination of others of the traditional
types, e.g. ritualist, scripturalist, primitivist.

That argument does not, however, find a place in these essays. My aim
here is a systematic study of the formal properties of scriptural authority,
as merely one (though possibly the major one) factor contributing to the
emergence of an independent and self-conscious religious community.
The literary uses, and hence communal functions, of scripture might be
(roughly) isolated as four: polemic, liturgical, didactic, and juridical, in
descending order of importance and (approximate) chronological order of
appearance. I believe that this set of priorities can be demonstrated from
the Muslim exegetical literature, examination of which constitutes half the
bulk of my book. The material adduced is intended to represent a cross-
section of Quranic commentary prior to the monumental work of Tabarl
(d. 923). Save for a kind of philosophical exegesis belonging to a later
period, the typology set out here includes the principal lines of inquiry
applied to the text of Muslim scripture: haggadic, halakhic, masoretic,
rhetorical, and allegorical. The manner in which the concept of authority
was progressively articulated by means of these exegetical types is the
formative principle and the purpose of my exposition, though it is of
course possible to interpret the evidence differently. It is, on the other
hand, quite impossible to mistake the chronology of the sources or to

PREFACE

xxiii

ignore the presence of Nachdichttmg in traditionist literary forms. Tra-
dition implies, and actively involves, historicization, and the growth of a
polemical motif into a historical fact is a process hardly requiring demon-
stration, e.g. the patriarchal narratives of the Old Testament. History, like
poetry, is mimetic and produces as many necessary truths as it contains
fortuitous facts (Lessing).

Pressed into the service of salvation history these two categories tend to
coalesce: everything becomes relevant, the most insignificant scraps of
information fall into place as components of the grand design. And it is
that design which scriptural exegesis is intended to illustrate, by recourse
to a range of standard hermeneutical techniques, all of ancient lineage and
indisputable merit. Islamic adaptation of these was a fairly uncomplicated
process of direct appropriation, and thus its analysis contains no surprises.
It can, on the other hand, be argued that the data of the theodicy (as set out
in Chapter I) limited from the outset its historical development: the
expression of communal purpose and authority could hardly have been
different from what it became (Weber). Sectarian forms of Islam reflect an
equally predictable pattern of divergence from what became the normative
(Sunni) expression. Again, the problem of diffusion and polygenesis. Some
aid towards a solution may be sought in the lexicon of exegesis, where
terminological caiques could, and probably do, indicate a community of
scholarship. I have examined these, but am unable to identify one single
path of diffusion. The marginalia of Judaeo-Christian history, what might
be called 'the sectarian milieu', await systematic exposition. That the wait
promises to be a long one is most persuasively illustrated in the quite
extraordinary literature generated by discovery of the Dead Sea Scrolls.
Comforted by speculative scholarship of such quality and quantity, I feel
no special compulsion to apologize for the conjectural nature of my own
efforts to depict the origins of Islam.

The final draft of this work was completed in July 1972, and I have thus
not taken account of studies published since that date (and quite possibly of
several published before that date). The sources employed in Chapter IV
are in part still in manuscript, and I am pleased to acknowledge the help-
fulness of the staffs of the several libraries in Istanbul listed in the Biblio-
graphy and of the British Museum. I should also like to express here my
gratitude to Simon Hopkins, who read and commented upon the final
version of these studies, having been for many years as my student exposed
to the cruder formulations of my earliest thoughts on the subject. I think
he is not satisfied with the result, but then, neither am I. Finally, I wish to
thank the School of Oriental and African Studies for granting me leave to
work in Istanbul, for including this book in the London Oriental Series,
and for meeting the expense of publication. j

July 1975

ABBREVIATIONS

AJSLL

American Journal of Semitic Languages and Literatures

AO

Acta Orientalia

BO

Bibliotheca Orientalis

BSOAS

Bulletin of the School of Oriental and African Studies

BSS

T. NSldeke, Beitrdge zur Semitischen Sprachwissenschaft

CHB

Cambridge History of the Bible

EI

Encyclopaedia of Islam, second edition, Leiden 1960-

GAL

C. Brockelmann, Geschichte der Arabischen Literatur

GAS

F. Sezgin, Geschichte des Arabischen Schrifttums

GdQ

T. Noldeke and F. Schwally, Geschichte des Qorans i-ii, G. Bergstrasser

and 0. Pretzl, Geschichte des Qorans iii

GVG

C. Brockelmann, Grundriss der Verglekhenden Grammatik der Semiti-

schen Sprachen

HUCA

Hebrew Union College Annual

IC

Islamic Culture

IQ

Islamic Quarterly

JA

Journal Asiatique

JAOS

Journal of the American Oriental Society

JNES

Journal of Near Eastern Studies

JQR

Jewish Quarterly Review

JRAS

Journal of the Royal Asiatic Society

JSS

Journal of Semitic Studies

MSOS

Mitteilungen des Seminars fur Orientalische Sprachen

MW

Muslim World

NBSS

T. Noldeke, Neue Beitrdge zur Semitischen Sprachwissenschaft

OLZ

Orientalistische Literaturzeitung

PAAJR

Proceedings of the American Academy for Jewish Research

REI

Revue des Etudes Islamiques

REJ

Revue des Etudes Juives

RO

Rocznik Orientalistyczny

SALP

N. Abbott, Studies in Arabic Literary Papyri

SI

Studia Islamica

VT

Vetus Testamentum

WI

Die Welt des Islams

WZKM

Wiener Zeitschrift fur die Kunde des Morgenlandes

ZA

Zeitschrift fur Assyriologie

ZDMG

Zeitschrift der Deutschen Morgenlandischen Gesellschaft

zs

Zeitschrift fur Semitistik

BIBLIOGRAPHY

Abbott, N. SALP I : Historical Texts, Oriental Institute Publications Ixxv, Univer-
sity of Chicago, 1957

SALP II : Quranic Commentary and Tradition, Oriental Institute Publica-
tions lxxvi, University of Chicago, 1967. See Wansbrough, BSOAS xxxi (1968)
613-16

'Abd al-Baql, M. Mujam gharib al-Qurdn, Cairo, 1950

'Abd al-Jabbar. Tanzih al-Qur'an, Beirut, n.d.

Tathbtt dalail al-nubuumia, MS. Shehid AH Pasha 1575

Abu 'Ubayd. Kitab al-nasikh vial-mansukh, MS. Ahmet III, 143

See A. Spitaler, 'Ein Kapitel aus den Fada'il al-Qur'an von Abu 'Ubaid

al-Qasim ibn Sallam', in Documenta Islamica Inedita, Berlin, 1953, 1-24

Abu 'Ubayda. Majdz al-Quran, Cairo, 1954-62

Ahrens, K. 'Christliches im Qoran', ZDMG lxxxiv (1930) 15-68, 148-90

Allard, M. 'Une Methode nouvelle pour l'etude du Coran, SI xvi (1961) 5-21

• Analyse conceptuelle du Coran sur cartes perforees, Paris, 1963

Altmann, A. 'Saadya's theory of revelation : its origin and background', in Saadya
Studies, Manchester, 1943, 4-25

Andrae, T. Die person Muhammeds in lehre und glauben seiner gemeinde, Stockholm,
1918

Ash'ari. Kitab maqdlat al-islamiyyin, Bibliotheca Islamica I, Wiesbaden, 1963
Asin Palacios, M. La Escatologia Musulmana en la Divina Comedia, Madrid, 1961.
Auerbach, E. Mimesis, Bern, 1967

Augapfel, J. 'Das Kitab im Qur'an', WZKM xxix (1915) 384-93
Bacher, W. 'Die Anfange der hebraischen Grammatik', ZDMG xlix (1895) 1-62,
335-93

Die exegetische Terminologie der judischen Traditionsliteratur i-ii, Leipzig,

1 899-1 905

Baghdad!, 'Abd al-Qahir. Kitab al-nasikh wal-mansukh, MS. Beyazit Umumi 445

Baljon, J. 'The "Amr of God" in the Qur'an', AO xxiii (1958) 7-18

Baqillani. I'jaz al-Qur'an, Cairo, 1963

Barr, J. The Semantics of Biblical Language, Oxford, 1961

Comparative Philology and the Text of the Old Testament, Oxford, 1968

Barth, J. 'Studien zur Kritik und Exegese des Qorans', Der Islam vi (1915-16)
113-48

Baydawi. Anwar al-tanzil, Bulaq, 1330

Beck, E. 'Die Kodizesvarianten der Amsar', Orientalia xvi (1947) 353-76

'Die Gestalt des Abraham am Wendepunkt der Entwicklung Muhammeds',

Le Musion lxv (1952) 73-94

xxviii

BIBLIOGRAPHY

Becker, C. H. 'Der Islam als Problem', Der Islam i (1910) 1-21 (Islamstudien I,
Leipzig, 1924, 1-23)

'Christliche Polemik und Islamische Dogmenbildung', ZA xxvi (191 1) 175-

95 (Islamstudien I, 432—49)

'Ubi sunt qui ante nos in mundo fuere', in Festschrift E. Kuhn, Breslau, 1916

87-105 (Islamstudien I, 501-19)

'Prinzipielles zu Lammens Sirastudien', Der Islam iv (1913) 263-9 (Islam-
studien I, 520-7 : 'Grundsatzlich.es zur Leben-Muhammed-Forschung')

Beitrage zur Geschichte Agyptens unter dem Islam i-ii, Strassburg, 1902-3

Bell, R. Bell's Introduction to the Quran, ed. W. M. Watt, Edinburgh, 1970

Ben-Zvi, I. 'Les Origines de Petablissement des tribus d'Israel en Arabie', he
Museon lxxiv (1961) 143-90

Bergstrasser, G. Einfuhrung in die semitischen Sprachen, Miinchen, 1928

Hebraische Grammatik i, Leipzig, 191 8

'Die Koranlesung des Hasan von Basra', Islamica ii (1926) 11-57

'Anfange und Charakter des juristischen Denkens im Islam', Der Islam xiv

(1925) 76-81

and Pretzl, O. Geschichte des Qorans III: Die Geschichte des Korantexts,

Leipzig, 1938

Bevan, A. 'Mohammed's Ascension', in Studien zur Semitischen Philologie und

Religionsgeschichte jfulius Wellhausen . . . gewidmet, Giessen, 1914, 49-61
Birkeland, H. Altarabische Pausalformen, Oslo, 1940

The Legend of the Opening of Muhammed's Breast, Oslo, 1955

Old Muslim Opposition against Interpretation of the Koran, Oslo, 1955

The Lord Guideth: Studies on Primitive Islam, Oslo, 1956

'The interpretation of Surah 107', SI ix (1958) 13-29

Bjorkman, W. Beitrage zur Geschichte der Staatskanzlei im islamischen Agypten,

Hamburg, 1928
Blachere, R. Introduction au Coran, Paris, 1959

Histoire de la litterature arabe i-iii, Paris, 1952-66

Black, M. An Aramaic approach to the Gospels and Acts, Oxford, 1967

Blanc, H. Communal Dialects in Baghdad, Cambridge, Mass., 1964

Blau, J. A Grammar of Christian Arabic, based mainly on South Palestinian Texts

from the first Millennium, CSCO Subsidia 27-9, Louvain, 1966-7. See Wans-

brough, BSOAS xxxi (1968) 610-13

On Pseudo-corrections in some Semitic Languages, Jerusalem, 1970

The Emergence and Linguistic Background of Judaeo- Arabic, Oxford, 1965

'The role of the bedouins as arbiters in linguistic questions and the Mas'ala

Az-Zunburiyya', JSS viii (1963) 42-51
'Some problems of the formation of the old Semitic languages in the light

of Arabic dialects', in Proceedings of the International Conference on Semitic

Studies, Jerusalem, 1965

'L' Apparition du type linguistique neo-arabe', REIxxxvi (1969) 191— 201

'Judaeo-Arabic in its linguistic setting', PAAJR xxxvi (1968) 1-12

'On the problem of the synthetic character of Classical Arabic as against

Judaeo-Arabic (Middle Arabic)', JQR lxiii (1972) 29-38

BIBLIOGRAPHY

xxix

Bloch, A. Vers und Sprache im Altarabischen, Acta Tropica Supplementum 5,
Basel, 1946

Bonebakker, S. Some Early Definitions of the Tawriya and Safadi's 'Fadd al-
Xitam 'an at-tawriya wa'l-istixddm', The Hague-Paris, 1966

Bouman, J. Le Conflit autour du Cor an et la solution d'al-Baqillant, Amsterdam, 1959

'The Doctrine of 'Abd al-Djabbar on the Qur'an as the created word of

Allah', in Verbum: Essays . . . Dedicated to Dr. H. W. Obbink, Utrecht, 1964,
67-86

Braunlich, E. 'Versuch einer literargeschichtlichen Betrachtungsweise altarabischer

Poesien', Der Islam xxiv (1937) 201-69
'Zur Frage der Echtheit der altarabischen Poesie', OLZ (1926) 825-33

Bravmann, M. 'On the spiritual background of early Islam and the history of its

principal concepts', Le Museon lxiv (1951) 317-56
Brockelmann, C. Geschichte der arabischen Literatur, Leiden, 1937-49

Grundriss der vergleichenden Grammatik der Semitischen Sprachen I— II, Berlin,

1908-13

Buhl, F. 'Uber Vergleichungen und Gleichnisse im Qur'an', AO ii (i924)_i-ii
'Zur Kuranexegese', AO iii (1925) 97-108

'Ein paar Beitrage zur Kritik der Geschichte MuhamrnedY, in Orientalische

Studien Theodor No'ldeke . . . gewidmet, Giessen, 1906, i, 7-22
Bukhari. Sahih iii, Leiden, 1868

Burton, J. 'Those are the high-flying cranes', JSS xv (1970) 246-65
'The collection of the Qur'an', Glasgow University Oriental Society, Trans-
actions xxiii, 1969-70 (1972) 42-60
Caetani, L. Anrtali deW Islam I, Milan, 1905

Caskel, W. 'Aijam al-'Arab: Studien zur altarabischen Epik', Islamica iii Suppl.
(1931) 1-99

'Zur Beduinisierung Arabiens', ZDMG ciii (1953) *28-36*

Cerulli, E. II 'Libro delta Scala' e la questione dellefonti arabo-spagnole delta Divina

Commedia, Studi e Testi 150, Citta del Vaticano, 1949
Chapira, B. 'Legendes bibliques attributes a Ka'b el-Ahbar', REJ Ixix (1919) 86-

107, lxx (1920) 37-43
Cohen, A. 'Arabisms in Rabbinic literature', JQ R iii (1912-13) 321—33
Cohen, D. 'Koine, langues communes et dialectes arabes', Arabica ix (1962) 1 19-44

Corre, A. 'Phonemic problems in the Masora', in Essays Presented to Chief Rabbi

Israel Brodie, London, 1968, 59-66
Corriente, F. 'On the functional yield of some synthetic devices in Arabic and

Semitic morphology', JQR lxii (1971) 20-50
Culley, R. Oral Formulaic Language in the Biblical Psalms, Toronto, 1967
Curtius, E. Europaische Literatur und Lateinisches Mittelalter, Bern, 1948
Danl. Kitab al-taysirfi 'l-qiradt al-sdb', Istanbul, 1930

Daube, D. 'Rabbinic methods of interpretation and Hellenistic rhetoric', HUCA
xxii (1949) 230-64

DeGoeje, M. 'Die Berufung Mohammed's', in Orientalische Studien Theodor
Noldeke . . .gewidmet, Giessen, 1906, i, 1-5

XXX

BIBLIOGRAPHY

Derenbourg, J. 'Quelques remarques sur la declinaison arabe', jfA iv (1844)
209-20

Derrett, J. Law in the New Testament, London, 1970
Dinawarl. Tafsir al-wadih, MS. Ayasofya 221-2
Dozy, R. Supplement aux Dictionnaires Arabes, Leiden, 1967
Eissfeldt, O. Einleitung in das Alte Testament, Tubingen, 1956
Elbogen, I. Der jiidiscke Gottesdienst in seiner geschichtlichen Entwicklung, Frank-
furt, 193 1

Farra'. Ma'anVl-Quran, Cairo, 1955-, and MS. Nurosmaniye 459
Ferguson, C. 'The Arabic Koine', Language xxxv (1959) 616-30
Fischer, A. 'Arab, aysh', ZDMG lix (1905) 807-18
'Usaiiid, usw.', ZDMG lvi (1902) 581

'Eine Qoran-Interpolation', in Orientalische Studien Theodor Nsldeke . . .

gewidmet, 1906, i, 33-55
'Die Mas'ala Zunburija', in A Volume of Oriental Studies Presented to E. G.

Browne, Cambridge, 1922, 150-6
Fischer, W. 'Ein Stuck vorklassischer, arabischer Kunstprosa in der Umm

Ma'bad-Legende', in Festschrift W. Eilers, Wiesbaden, 1967, 318-24

'Silbenstruktur und Vokalismus im Arabischen', ZDMG cxvii (1967)

30-77

Frye, N. Anatomy of Criticism, Princeton, 1957

Fuck, J. ' Arabiya; recherches sur Vhistoire de la langue et du style arabe, tr. C.

Denizeau, Paris, 1955
Galliner, S. Saadia Al-fajjuml's arabische Psalmeniibersetzung und Commentar

(Psalm 73-89), Berlin, 1903
Garbell, I. 'Remarks on the historical phonology of an East Mediterranean Arabic

dialect', Word xiv (1958) 303-37
Geiger, A. Was hat Mohammed aus dem Judenthume aufgenommen? , Leipzig, 1902

Gerhardsson, B. Memory and Manuscript: Oral Tradition and Written Transmission
in Rabbinic Judaism and Early Christianity, Copenhagen, 1964

Gertner, M. 'The Masorah and the Levites : an essay in the history of a concept',
VT x (i960) 241^72

'Terms of scriptural interpretation : a study in Hebrew semantics', BSOAS

(1962) 1-27

'The terms pharisaioi, gazarenoi, hupokritai : their semantic complexity and

conceptual correlation', BSOAS xxvi (1963) 245-58
'On translating medieval Hebrew writing', JRAS (1963) 163-93

Goitein, S. 'Ramadan, the Muslim month of fasting', in Studies in Islamic History

and Institutions, Leiden, 1966, 90-110
'A turning-point in the history of the Muslim state', IC xxiii (1949) 120-35

(Studies 149-67)

'The Birth-hour of Muslim law', MW 1 (i960) 23-9 (Studies 126-34)

Goldziher, I. Abhandlungen zur Arabischen Philologie i-ii, Leiden, 1896-99

Die Richtungen der islamischen Koranauslegung, Leiden, 1920

Muhammedanische Studien i-ii, Halle, 1889-90

BIBLIOGRAPHY

xxxi

Die Zahiriten, Leipzig, 1884

Vorlesurtgen iiber den Islam, Heidelberg, 1910

Gesammelte Schriften i-v, ed. J. Desomogyi,--Hildesheim, 1967-70 (GS)

— 'Beitrage zur Geschichte der Sprachgelehrsamkeit bei den Arabern I— III',
SKAW lxvii (1871) 207-51, lxxii (1872) 587-631, lxxiii (1873) 511-52 (GS i,
7-51, 9I-I35. 187-228)

'Linguistisches aus der Literatur der muhammedanischen Mystik', ZDMG

xxvi (1872) 764-85 {GS i, 165-86)
'Uber muhammedanische Polemik gegen Ahl al-kitab', ZDMG xxxii (1878)

341-87 (GSii,i- 4 7)
'Das Princip des Istishab in der muhammedanischen Gesetzwissenschaft',

WZKM i (1887) 228-36 (GS ii, 182-90)

— 'Die Ginnen der Dichter', ZDMG xlv (1891) 685-90 (GS ii, 400-5)

— 'Der Chatib bei den alten Arabern', WZKM vi (1 892) 97-102 (GS iii, 27-32)

'Usages juifs d'apres la litterature religieuse des Musulmans', REJ xxviii

(1804) 75-94 (GS iii, 322-41)
'Neue Materialien zur Litteratur des Uberlieferungswesens bei den Muham-

medanern', ZDMG 1 (1896) 465-506 (GS iv, 69-110)

'Melanges judeo-arabes', REJ xliii (1901) 1-14 (GS iv, 263-76)

'Neutestamentliche Elemente in der Traditionslitteratur des Islam', Oriens

Christiamts (1902) 390-7 (GS iv, 315-22)
'Die Bedeutung der Nachmittagszeit im Islam', Arckiv ftir Religionstvissen-

schaft ix (1906) 294-302 (GS v, 23-31)
'Aus der Theologie des Fachr al-din al-Razi', Der Islam iii (19 12) 213-47

(GS v, 237-71)

'Bemerkungen zur neuhebraischen Poesie', JQR xiv (1902) 719-36

'Kampfe urn die Stellung des Hadit im Islam', ZDMG lxi (1907) 860-^72

(GS v, 86-98)

Gotz, M. 'MaturidI und sein Kitab Ta'wllat al-Qur'an", Der Islam xli (1965)
27-70

Grohmann, A. Einfuhrung und Chrestomathie zur arabischen Papyruskunde,
Prague, 1955

Grunebaum, G. von. A Tenth-century Document of Arabic literary Theory and

Criticism, Chicago, 1950
Kritik und Dichtkunst: Studien zur arabischen Liter atur geschichte, Wiesbaden,

1955

Grunbaum, M. Neue Beitrage zur semitischen Sagenkunde, Leipzig, 1893
Guidi, M. La lotta tra V Islam e il Manicheismo, Rome, 1927
Hasan, A. 'The theory of naskh', Islamic Studies iv (1965) 181-200

Heinrichs, W. Arabische Dichtung und Griechische Poetik, Beirut, 1969. See Wans-

brough, BSOAS xxxiii (1970) 615-17
Heller, B. 'Recits et personnages bibliques dans la legende mahometane', REJ

lxxxv (1928) 113—36

'La Legende biblique dans l'lslam', .REJ'xcviii (1934) 1-18

Hibatallah. Kitab al-nasikh loal-mansiikh, Cairo, i960

xxxii

BIBLIOGRAPHY

Hirschberg, J. Jiidische und christliche Lehren im vor- und fruhislamischen Arabien,
Cracow, 1939

Hirschfeld, H. 'Historical and legendary controversies between Mohammed and
the Rabbis', JQR x (1897-8) 100-16

'Mohammedan criticism of the Bible', JQR xiii (1900-1) 222-40

New Researches into the Composition and Exegesis of the Qoran, London, 1902

Horovitz, J. Koranische Untersuchungen, Berlin-Leipzig, 1926

'Zur Muhammadlegende', Der Islam v (1915) 41-53

'Muhammeds Himmelfahrt', Der Islam ix (1919) 159-83

'Biblische Nachwirkungen in der Sira', Der Islam xii (1922) 184-9

Horst, H. 'Zur Oberlieferung im Korankommentar at-Tabaris', ZDMG ciii (1953)
290-307

Husayn, Taha. 'La Rhetorique arabe de Djahiz a 'Abd al-^ahir', in rCudama b.

Dja'far: Nakd an-nathr, Cairo, 1933, 3-24
Hyman, A. 'Maimonides' "Thirteen Principles'", Jewish Medieval and Renaissance

Studies, Cambridge, Mass., 1967, 119-44
Ibn 'Abbas. Kitab baydn lughdt al-Qur'dn, MS. Esad Efendi 91
Kitab gharlb al-Qur'an, MS. Atif Efendi 2815

'Masd'il Ndfi' b. Azraq' : see 'Abd al-Baql, Mu'jam gharlb al- Quran, 234-92

Ibn 'Abd Rabbih. Al-'Iqd al-jarid, Bulaq, 1293
Ibn Abl '1-Isba'. Badx al-Qur'dn, Cairo, 1957
Ibn Anbari. Al-Insdffi masd'il al-kkildf, Cairo, 1961
Ibn 'Arab!. Ahkdm al-Qur'dn, Cairo, 133 1

Sharh to Tirmidhl, $ahih, Cairo, 1934

Ibn Athir. Kitab al-kdmil, Leiden, 1851-76
Ibn Hazm. Kitab al-fisal zoal-milal, Cairo, 1321
Ibn Hisham. See Ibn Ishaq

Ibn Hisham al-Ansari. Mugkni 'l-labib, Cairo, 1969

Ibn Ishaq. Al-Sira al-nabavnyya (Ibn Hisham), Cairo, 1955

Ibn Kalbl, Hisham. Kitab al-Asndm, Cairo, 1924

Ibn Kammuna (Sa'd b. Mansur). Taiufih al-abhdth lil-milal al-thaldth, Berkeley-
Los Angeles, 1967

Ibn Kathir. Tafstr, Beirut, 1966

Ibn Munayyir. 'Al-Intisaf ' : sharh to Zamakhshari, Al-Kaskshdf, Beirut, 1967
Ibn Mu'tazz. Kitab al-Badl' , London, 1935

Ibn Muqaffa'. 'Risalat fil-Sahaba', in M. Kurd 'All, Rasd'il al-bulagha, Cairo,

1946, 117-34
Ibn Qutayba. Ta'tml Mushkil al-Qur'dn, Cairo, 1954
Ibn Sa'd. fabaqat, Cairo, 1939

Ibn Wahb. Al-Burhdn fi wujuh al-bayan, Baghdad, 1967
Ibn Abi Dawud. See JefTery, Materials

Isfahan!, Raghib. Hall mutashdbihdt al-Qur'dn, MS. Ragip Pasha 180
Izutsu, T. Ethico-religious concepts in the Quran, Montreal, 1966
Jahiz. 'Kitab hujaj al-nubuwwa*, in Rasd'il, Cairo, 1933, 117-54

BIBLIOGRAPHY

xxxiii

Al-Bayan teal-tabyin i-iii, Cairo, 1947

'Kitab al-tarbi' wal-tadwir', in Tria Opuscula, Leiden, 1903, 86—107

Ja'far al-§adiq. Tafsir, MS. Nafiz Pasha 65
Jassas. Ahkam al-Quran, Istanbul, 1335-8
Jastrow, M. A Dictionary of the Targwnim, New York, 1950

Jeffery, A. Materials for the history of the text of the Quran (inc. Ibn Abi Dawud,

Kitab al-masahif), Leiden, 1937

'The Qur'an as scripture', MW xl (1930) 41-55, 106-34, 185-206, 257-75

Jensen, P. 'Das Leben Muhammeds und die David-Sage', Der Islam xii (1922)

84-97

Johnson, A. 'Mashal', in Wisdom in Israel and in the ancient Near East, Suppl. VT
iii (i960) 162-9

Jones, J. 'The chronology of the maghazi — a textual survey', BSOAS xix
(1957) 245-80

Jullandri, R. 'Qur'anic exegesis and classical tafsir', IQ xii (1968) 71-1 19
Jurjani. Asrar al-baldgha, Istanbul, 1954
Dalaili'jdz al-Quran, Cairo, 1372

Kahle, P. 'The Arabic readers of the Koran', JNES viii (1949) 65-71
Kalbl, Muhammad. Tafsir, MS. Ayasofya 118, Hamidiye 40
Katsch, A. Judaism in Islam, New York, 1954

Khoury, A. Les Theologiens byzantins et V Islam, Louvain-Paris, 1969. SeeWans-

brough, BSOAS xxxiii (1970) 391-3
Khurasani, 'Afa' (Pseudo). Tafsir (ikhtilaf), MS. Ahmet III, 310
Kisa'i. Kitab Mushtabihdt al-Quran, MS. Beyazit Umumi 436
Kister, M. 'You shall only set out for three mosques' : a study of an early tradition',

Le Museon lxxxii (1969) 173-96

'A bag of meat: a study of an early hadith', BSOAS xxxiii (1970) 267-75

— — 'Al-tahannuth: an inquiry into the meaning of a term', BSOAS xxxi (1968)

223-36

'Al-hira: some notes on its relations with Arabia', Arabica xv (1968) 143-69

Koch, K. The Growth of the Biblical Tradition, London, 1969

Kofler, H. 'Reste altarabischer Dialekte', WZKM xlvii (1940) 61-130, 233-62,

xlviii (1941) 52-88, 247-74, xlix (1942) 15-30, 234-56
Kluge, F. and Gotze, A. Etymologisches W&rterbuch der deutschen Sprache, Berlin

1953

Kopf, L. 'Religious influences on medieval Arabic philology', SI v (1956) 33-59
'The treatment of foreign words in medieval Arabic lexicology', Scripta

Hierosolymitana ix (1961) 191-205
Krauss, S. 'Talmudische Nachrichten iiber Arabien', ZDMG lxx (1916) 32i~53»

Ixxi (1917) 268-9

KUnstlinger, D. ' "Kitab" und "ahlu 1-kitabi" im Kuran', RO iv (1926) 238-47

' "Islam", "Muslim", "Aslama" im Kuran'. RO xi (1935) 128-37

'Die Namen der Gottes-Schriften im Quran', RO xiii (1937) 72-84

« 'Uzair ist der Sohn Allah's', OLZ (1932) 381-3

'Sab'an min al-mathani', OLZ (1937) 596-8

xxxiv

BIBLIOGRAPHY

Lammens, H. U Arable occidentale avant Vhegire, Beirut, 1928

Lauterbach, J. Saadja Al-fajjumi's arabische Psalmeniibersetzung. und Commentar

(Psalm 107-24), Berlin, 1903
Lausberg, H. Handbuch der literarischen Khetorik, Miinchen, i960
Lehmann, E., and Pedersen, J. 'Der Beweis fiir die Auferstehung im Koran, Der

Islam v (1914) 54-61
Lidzbarski, M. 'Salam und Islam', ZS i (1922) 85-96

Lindblom, J. 'Wisdom in the Old Testament prophets', in Wisdom in Israel and in

the ancient Near East, Suppl. PT iii (i960) 192-204
Loewe, R. 'The "Plain" meaning of scripture in early Jewish exegesis', in Papers

of the Institute of Jewish Studies, Jerusalem, 1965, 140-85
'The medieval history of the Latin Vulgate', CHB ii, Cambridge, 1969,

102-54

Ma'arri. Risdlat al-ghufran, Cairo, 1950

Anon. Kitdb al-Mabani, Cairo, 1954 (Jeffery, Two Muqaddimas)

Malati. Kitab al-tanbth wal-radd 'aid ahl al-ahwa wal-bida', Bibliotheca Islamica

IX, Istanbul, 1936
Maimonides. Dalalat al-hairtn i-iii, Paris, 1856-66 (S. Munk)

Iggeret Teman, New York, 1952 (A. Halkin)

Malik b. Anas. Al-Muwatta', Cairo, 1951

Margoliouth, D. 'Omar's instructions to the kadi', JRAS (1910) 307-26
'The origins of Arabic poetry', JRAS (1925) 417-49

Massignon, L. RecueH de textes inedits concernant I'kistoire de la mystique en pays

d'Islam i, Paris, 1929
Masson, D. Le Coran et la revelation jude'o-chretienne: etudes comparies, Paris,

1958

Mann, J. The Jews in Egypt and in Palestine under the Fatimids i-ii, Oxford,
1920-2

Maturidi. Kitab Ta'icilat al-Qur'dn, MS. Medine 179, 180
Mehren, A. von. Die Rhetorik der Araber, Kopenhagen-Wien, 1853
Metzger, B. The Text of the New Testament, Oxford, 1968

Mirsky, A. 'Biblical variants in medieval Hebrew poetry', Textus iii (1963) 159-62

Mittwoch, E. 'Die Berliner arabische Handschrift Ahlwardt No. 683', in A Volume
of Oriental Studies Presented to E. G. Browne, Cambridge, 192a, 339-44

Monroe, J. 'Oral composition in Pre- Islamic poetry', Journal of Arabic Literature
iii (1972) 1-53

Moubarac, Y. Abraham dans le Coran, Paris, 1958

Moscati, S. (ed.). An Introduction to the Comparative Grammar of the Semitic Lan-
guages, Wiesbaden, 1964

(ed.). Linguistxca Semitica; presente efuturo, Studi Semitici 4, Rome, 1961

Mowinckel, S. 'Psalm criticism between 1900 and 1935', VT v (1955) 13-33
Muilenburg, J. 'A study in Hebrew rhetoric: repetition and style', Suppl. VTi
(1953) 97~" 1

Miiller, F. Untersuchungen zur Reimprosa im Koran, Bonn, 1969. See Wansbrough,
BSOAS xxxiii (1970) 389-91

BIBLIOGRAPHY

XXXV

Murtada, Sharif. Amdli, Cairo, 1954

Muqatil b. Sulayman. Tafsir, MS. Hasan Husnu 17

Kitab tafsir al-khamsmi'at aya min al-Qutdn, MS. British Museum Or.

6333

Kitab al-Ashbdh zval-nazir fi tafsir al-Qurdn, MS. Beyazit Umumi 561

Muslim. Sahih viii, Cairo, 1332

Nahhas. Kitab al-ndsikh wal-mansukh, Cairo, 1038

xxxi . .
Noldeke, T. Beitrage zur Kermtnis der Poesie der alten Araber, Hannover, 1864

Delectus Veterum Carminum Arabicorum, Berlin 1890

Zur Grammatik des classischen Arabisch (ed. Spitaler), Darmstadt, 1963

Beitrage zur semitischen Sprachwissenschaft, Strassburg, 1904

Neue Beitrage zur semitischen Sprachwissenschaft, Strassburg, 19 10

and Schwally, F. Geschichte des Qorans i-ii, Leipzig, 1909-19. See Berg-

strasser-Pretzl
Norden, E. Die Antike Kunstprosa, Stuttgart, 1958

Norris, H. 'New evidence on the life of 'Abdullah B. Yasin and the origins of the
the Almoravid movement', Journal of African History xii (1971) 253-68

Nyberg, H. 'Zum Kampf zwischen Islam und Manichaismus', OLZ xxxii (1929)
425-41

Obermann, J. 'Political theology in early Islam: Hasan al-Basri's treatise on qadar',

JAOS lv (1935) 138-62

'Koran and Agada: the events at Mount Sinai', AJSLL lvii (1941) 23-48

' Islamic origins : a study in background and foundation', in The Arab Heritage,

Princeton, 1944, 58-120
Ory, S. 'Un Nouveau Type de mushaf : Inventaire des Corans en rouleaux de

provenance damascaine, conserves a Istanbul', REI xxxiii (1965) 87—149

O'Shaughnessy, T. The Koranic Concept of the Word of God, Rome, 1948

'The seven names for Hell in the Qur'an', BSOAS xxiv (1961) 444-69

Muhammad's Thoughts on Death: a Thematic Study of the Quranic Data,

Leiden, 1969. See Wansbrough, BSOAS xxxiii (1970) 613-15

Paret, R. 'Der Koran als Geschichtsquelle', Der Islam xxxvii (1961) 26-42

Der Koran: Kommentar und Konkordanz, Stuttgart, 1971

Pedersen, J. Israel i-iv, London-Copenhagen, 1926-40

'The Islamic preacher: wa'iz, mudhakkir, qass', in Goldziher Memoria

Volume i, Budapest, 1948, 226-51

'The criticism of the Islamic preacher', WI ii (1953) 215-31

See Lehmann, E.

Perlmann, M. 'Another Ka'b al-Ahbar story', JQR xlv (1954) 48-58

Qarfajanni. See Heinrichs, W., Arabische Dichtung

Qalqashandi. Subh al-a'shd, Cairo, 1914-20

Qasim b. Ibrahun. Kitab al-radd 'aid 'l-zindiq (see Guidi, M.)

QazwM. Skuriih al-talkhis i-iv, Cairo, 1937

Qummi. Tafsir, Najaf, 1386

Qurfubl. Al-Jdmt 'li-ahkdm al-Qurdn, Cairo, 1967

xxxvi

BIBLIOGRAPHY

Rabin, C. Ancient West- Arabian, London, 1951

"The beginnings of Classical Arabic', SI iv (1955) 19-37

Qumran Studies, Oxford, 1957

'The historical background of Qumran Hebrew', Scripta Hierosolymitana iv

(1958/1965) 144-61
Rad, G. von. Old Testament Theology i-ii, Edinburgh, 1962-5
Reckendorf, H. Die Syntaktischen Verhd'ltnisse des Arabischen i-ii, Leiden,

1895-8

Arabische Syntax, Heidelberg, 1921

Reuschel, W. 'Wa-kana Uahu 'allman rahiman', in Studia Orientalia in Manor iam

Caroli Brockelmann, Halle, 1968, 147-53
Richter, W. Exegese als Literaturwissenschaft, Gfittingen, 197 1
Ritter, H. 'Studien zur Geschichte der islamischen Frommigkeit I. Hasan al-

Basri, Der Islam xxi (1933) 1-83
Rosenthal, F. 'The influence of the Biblical tradition on Muslim historiography',

in Historians of the Middle East, London, 1962, 35-45
Saadya. Kitdb al-Amdnat wal-i'tiqadat, Leiden, 1880

See Galliner, S., and Lauterbach, J.

Sabbagh, T. La Mitaphore dans le Coran, Paris, 1943

Schacht, J. The Origins of Muhammadan Jurisprudence, Oxford, 1953

An Introduction to Islamic Law, Oxford, 1964

Schapiro, I. Die haggadischen Elemente im erzahlenden Teil des Korans i, Leipzig,
1907

Schoeps, H. Theologie und Geschichte des Judenchristentums, Tubingen, 1949

Schreiner, M. 'Zur Geschichte der Polemik zwischen Juden und Muhammeda-

nern', ZDMG xlii (1888) 591-675
Schrieke, B. 'Die Himmelsreise Muhammeds', Der Islam vi (1915-16) 1-30
Schwarz, M. 'The Letter of al-Hasan al-Basrf , Oriens xx (1967/9) 15-30

Schwarzbaum, H. 'The Jewish and Moslem versions of some theodicy legends',
Fabula iii (1959-60) 119-69

Seeligmann, I. ' Voraussetzungen der Midraschexegese', Suppl. FTi(i953) 150-81
Segal, M. A Grammar of Mishnaic Hebrew, Oxford, 1927

Sellheim, R. 'Prophet, Caliph und Geschichte: Die Muhammad-Biographie des

Ibn Ishaq', Oriens xviii-xix (1965-6/67) 33-91
Semaan, K. 'Al-Nasikh wal-mansukh', IQ vi (1961) 11-29
Serjeant, R. "The "Constitution of Medina" ', IQ viii (1964) 3-16
Sezgin, F. Geschichte des Arabischen Schrifttums i, Leiden, 1967
Shafi'i. Al-Risala, Cairo, 1940
Ahkam al-Quran (Bayhaqi), Cairo, 195 1-2

Shahid, I. 'A contribution to Koranic exegesis', in Arabic and Islamic Studies in

Honor of Hamilton A. R. Gibb, Leiden, 1965, 563-80
Shahrastani. Kitdb al-Milal wal-nihal, Cairo, 1321

Sidersky, D. Les Origines des legendes musulmanes dans le Coran, Paris, 1933
Sister, M. 'Metaphern und Vergleiche im Koran' MSOS xxxiv (1931) 104-54

BIBLIOGRAPHY

xxxvii

Speyer, H. Die BibUschen Erzdhlungen im Qoran, Hildesheim, 1961
Spitaler, A. Review of Fiick, 'Arabiya, in BO x (1953) 144-50

Review of Bloch, Vers und Sprache, in OriensM (1949) 317-22

'Arabisch', in Linguhtica Semitica, 115-38 (see Moscati, S.)

See NSldeke, Zur Grammatik

Steinschneider, M. Polemische und apologetische Literatur in arabischer Sprache,
Leipzig, 1877

Stetter, E. Topoi und Schemata im hadit, Tubingen, 1965

Stieglecker, H. Die Glaubenslehren des Islam i-iv, Paderbom, 1959-62

Strack, H. Introduction to the Talmud and Midrash, Philadelphia, 1945

Subkl. Shuruh al-talkhls i-iv (see Qazwlni)

Sufyan Thawri. Tafsir al-Qur'an al-karim, Rampur, 1965

Suhayll. Al-Rawd al-unuf, Cairo, 1332

Suyirp. Al-Itqan fi'ulum al-Qur'dn, Cairo, 1967

Al-Muzhtr fi'ulum al-lugha, Cairo, 1958

and Mahalli. Tafsir al-Jalalayn, Damascus, 1385

Tabari. Annates, Leiden, 1 879-1 901
Tafsir, Cairo, 1321-8

Kitab Ikhtildf al-fuqahd' (J. Schacht), Leiden, 1933

Tahanawi. Kitab Kashshaf istildhdt al-funun, Calcutta, 1862
Tirmidhi. Sahih xi-xii, Cairo, 1934

Torrey, C. The Jewish Foundation of Islam, New York, 1967
Tustari. Tafsir al-Qur'an al-'azim, Cairo, 1329

Tyan, E. 'Methodologie et sources du droit en Islam', SI x (1959) 79-109
Ullendorff, E. 'Is Biblical Hebrew a language?', BSOAS xxxiv (1971) 241-55
Ullmann, M. Untersuchungen zur Ragazpoesie, Wiesbaden, 1966
Vajda, G. ']uifa et Musulmans selon le hadit', J A ccxxix (1937) 57-127
Vermes, G. Scripture and Tradition in Judaism, Leiden, 1961

"The use of bar nash/bar nasha in Jewish Aramaic', in Black, An Aramaic

Approach, 310-^28

'Bible and Midrash: early Old Testament exegesis', CHB i, Cambridge,

1970, 199-231

Vollers, K. 'Arabisch und Semitisch : Gedanken uber eine Revision der semitischen
Lautgesetze', ZA ix (1894) 165-217

• Review of Noldeke, Zur Grammatik, in ZA xii (1897) 125-39

Volkssprache und Schriftsprache im alten Arabien, Strassburg, 1906

Vossler, K. Einfuhrung ins Vulga'rlatein (H. Schmeck), Mttnchen, 1954
Wahddx. Kitab Asbdb al-nuzM, Cairo, 13 15

Wansbrough, J. 'A Note on Arabic rhetoric', in Lebende Antike: Symposion filr

Rudolf Siihnel, Berlin, 1967, 55-63

'Arabic rhetoric and Qur'anic exegesis', BSOAS xxxi (1968) 469-85

■ 'Majaz al-qur'an: periphrastic exegesis', BSOAS xxxiii (1970) 247-66

Wzqidi. Kitab al-Maghazi (M.Jones), London, 1966. See Wansbrough, BSOAS

xxxi (1968) i48- i 9

xxxviii

BIBLIOGRAPHY

Watt, W. 'Early discussions about the Qur'an', MWsl (1950) 27-40, 96-105

'The dating of the Qur'an: a review of Richard Bell's theories', JRAS (1957)

46-56

'The Materials used by Ibn Ishaq', in Historians of the Middle East, London,

1962, 23-34

Wehr, H. Review of Fuck, ' Arablya, in ZDMG cii (1952) 179-86
Weil, G. Die grammatischen Schulen von Kufa und Basra, Leiden, 1913

Grundriss und System der altarabischen Metren, Wiesbaden, 1958

Weinreich, U. Languages in contact, The Hague-Paris, 1968
Weisgerber, L. Deutsch als Volksname, Stuttgart, 1953

Weisweiler, M. ''Abdalqahir al-Curcani's Werk fiber die Unnachahmlichkeit des

Korans und seine syntaktisch-stilistischen Lehren, Oriens xi (1958) 77—121
Wellhausen, J. Reste arabischen Heidentutns, Berlin, 1927

Prolegomena zur altesten Geschichte des Islams, Skizzen und Vorarbeiten VI,

Berlin, 1899

— — Das arabische Reich und sein Stum, Berlin, 1902

Wensinck, A. 'Muhammed und die Propheten', AO ii (1924) 168-98

A Handbook of Early Muhammadan Tradition, Leiden, 1927

The Muslim Creed, Cambridge, 1932

Westermann, C. Grundfortnen prophetischer Rede, Munchen, 1968
Wetzstein, I. 'Sprachliches aus den Zeltlagern der syrischen Wiiste', ZDMG xxii
1868) 69-194

Widengren, G. Muhammad, the Apostle of God, and his Ascension, Uppsala-
Wiesbaden, 1955

'Oral tradition and written literature among the Hebrews in the light of

Arabic evidence, with special regard to prose narratives', AO xxiii (1959) 201-62
Wieder, N. Thejudean Scrolls and Karaism, London, 1962
Wild, S. Das Kitab al- Ain und ate arabische Lexikographie, Wiesbaden, 1965
Wurthwein, E. The Text of the Old Testament, Oxford, 1957
Zabldl. Taj al-'Arus, Cairo, 1306-7

Zajjaj. Kitab Trdb al-Qurdn wa-md amhi, MS. Ahmet III, 122-3, Nurosmaniye
US

Zamakhshari. Al-Kashshaf 'an ftaqa'iq al-tanzil i-iv, Beirut, 1967

Al-Mufassal, Christiana, 1879

Zuhri. Kitab Tanzil al- Quran, Beirut, 1963

Zunz, L. Die Gottesdienstlichen Vortrage der Juden, Frankfurt, 1892

MANUSCRIPTS UTILIZED IN QURANIC STUDIES

. 'Abd al-Jabbar, Tathbit dala'il al-nubuwwa

An edition was published in 1966, edited by Abd al-Kanm
'Uthman.

. Abu 'Ubayd, Kitab al-nasikh wal-mansukh

A facsimile edition of the manuscript Ahmet III 143 (209!?.)
was published by F. Sezgin, Frankfurt 1985. The manuscript was
edited with commentary by John Burton, Cambridge 1987; also
edited by Muhammad al-Mudayfir, Riyad 1990. See A. Rippin,
"Abu 'Ubaid's Kitab al-nasikh wa'l-mansukh," BSOAS 53 (1990):
319—20, for details on the manuscript fragment, Turk ve Islam
Eserli Miizesi 7892.

. Abu 'Ubayd, Fada'il al-Qur'an

Wansbrough employed an extract published by Spitaler. The
full text has been edited by Wahbi Sulayman Khawaji, Beirut 1991.

. ahBaghdadi, Abd al-Qahir, Kitab al-nasikh wal-mansukh

The text has been published, ed. Hilml Kamil As'ad Abd al-
Hadl, 'Amman 1987, based upon the manuscripts Berlin
Petermann 555 (a somewhat broken copy) and Beyazit 445. A
third significant manuscript exists, Manisa 178, written in the
eighth century, loiff. For further details on this text, see A.
Rippin, "The Exegetical Literature of Abrogation: Form and
Content," in Studies in Islamic and Middle Eastern Texts and
Traditions in Memory of Norman Calder (Oxford, 2000), pp.
213-31-

. Dlnawarl, Tafsir al-wadih

The text has been published under the title Tafsir ibn Wahb
musamma al- Wadih h tafsir al-Qur'an al-karim, ed. Ahmad Farid
(Beirut, 2003). However, the work ascribed to Kalbi — see number
9 below — and at the same time ascribed to Ibn 'Abbas and
al-Flruzabadl and printed many times under the title Tanwir al-
miqbas min > tafsir Ibn 'Abbas — is identical; see A. Rippin, " Tafsir

xl

MANUSCRIPTS UTILIZED

Ibn 'Abbas and Criteria for Dating Early tafsir Texts," Jerusalem
Studies in Arabic and Islam 18 (1994): 38-83 (= A. Rippin, The
Qur'an and Its Interpretative Tradition [Aldershot, UK, 2001 ],
chap. 15).

6. Ibn 'Abbas, Kitab bayan lughat al-Qur'an

Wansbrough indicates (p. 219) that the manuscript Esad
Efendi 91 with this title contains the text of Ibn 'Abbas, Gharib al-
Qur'an. For further clarification of the works under both these
titles, see A. Rippin, "Ibn 'Abbas's al-Lughat f-'l-Qur'an" BSOAS
44 (1981): 15-25; and A. Rippin, "Ibn Abbas's Gharib al-
Qur'an," BSOAS 46 (1983): 332-33 (= A. Rippin, The Qur'an
and Its Interpretative Tradition, chaps. 13 and 14). The work is
published (as indicated by Wansbrough. but not consulted by him)
as Kitab al-lughat fi'l-Qur'an rawaya ibn Hasnun al-Muqri bi-
isnadihi ila Ibn 'Abbas, ed. Salah al-Din al-Munajjid (Cairo, 1946;
reprint, Beirut, 1972).

7. Ibn 'Abbas, Kitab gharib al-Qur'an

The work is identical to Ibn 'Abbas, Kitab bayan lughat al-
Qur'an; see number 6 above.

8. Isfahan!, Raghib, Hall mutashabihat al-Qur'an

No further work appears to have been done on this
manuscript.

9. Ja'far al-Sadiq, Tafsir

No further work appears to have been done on this
manuscript. Meir M. Bar-Asher, Scripture and Exegesis in Early
Imami Shiism (Leiden, 1999), comments that such texts are
"clearly ahistorically ascribed" to the early Imams (p. 7).

10. Kalbl, Tafsir

Published many times; see number 5 above.

1 1. Khurasanl, Ata' (Pseudo), Tafsir (ikhtilaf)

No further work appears to have been done on these folios.

MANUSCRIPTS UTILIZED

xli

2. Kisa°I, Kitib mushtabihit al-Qur'an

Published under the title Kitib mutashibih al-Qur'an, ed.
Sablh al-Tamlml (Tarabulus, 1994).

3. al-Maturldl, Kitab ta'wilat al-Qur'in

Only vol. 1 has been published under the title Ta'wilat ahl al-
sunna, ed. Ibrahim and al-Sayyid 'Awadayn (Cairo, 1971); also, ed.
Muhammad Mustafid al-Rahman (Baghdad, 1983).

4. Muqatil ibn Sulayman, Tafsir

Published in four volumes (plus a fifth volume as in index and
study), ed. 'Abd Allah Mahmud Shihata (Cairo, 1980—89); also
published in three volumes, ed. Ahmad Farid (Beirut, 2003).

5. Muqatil ibn Sulayman, Kitab tafsir al-khamsmi'at aya min al-
Qur'an

Published, ed. Isaiah Goldfeld (Shfaram, 1 980).

6. Muqatil ibn Sulayman, Kitab al-ashbah wa'l-nazair fi tafsir al-
Qur'in

Published under the title Kitab al-ashbah wa'l-naza'Ir fi '1-
Qur'an, ed. c Abd Allah Mahmud Shihata (Cairo, 1975).

7. Zajjaj, Kitib irab al-Qur'in wa-ma'inihi

Published in five volumes under the title Ma'am '1-Qur'an wa-
irabuhu, ed. 'Abd al-Jalll *Abduh Shalabl (Beirut, 1988).

I

REVELATION AND CANON

I. THE DOCUMENT

Once separated from an extensive corpus of prophetical logia, the Islamic
revelation became scripture and in time, starting from the fact itself of
literary stabilization, was seen to contain a logical structure of its own. By
the very achievement of canonicity the document of revelation was
assured a kind of independence, both of historical traditions commonly
adduced to explain its existence and of external criteria recruited to
facilitate its understanding. But the elaborate and imposing edifice of
classical Quranic scholarship is hardly monolithic, and discernible lines of
cleavage correspond to the number of options left open to the most
fundamental lines of inquiry. Both formally and conceptually, Muslim
scripture drew upon a traditional stock of monotheistic imagery, which
may be described as schemata of revelation. Analysis of the Quranic
application of these shows that they have been adapted to the essentially
paraenetic character of that document, and that, for example, originally
narrative material was reduced almost invariably to a series of discrete and
parabolic utterances. An illustration is Surat Yusuf, often cited as a single
instance of complete and sustained narrative in the Qur'an. In fact,
without benefit of exegesis the Quranic story of Joseph is anything but
clear, a consequence in part of its elliptical presentation and in part of
occasional allusion to extra-Biblical tradition, e.g. verses 24, 67, 77.' It
may, indeed, be supposed that the public for whom Muslim scripture was
intended could be expected to supply the missing detail. A distinctly
referential, as contrasted with expository, style characterizes Quranic
treatment of most of what I have alluded to as schemata of revelation,
exhibited there as components of earlier established literary types. The
technique by which a theme is repeatedly signalled but seldom developed
may be observed from an examination in their Quranic form of those
themes traditionally associated with literature of prophetical expression.
Not merely the principal themes, but also the rhetorical conventions by
which they are linked and in which they are clothed, the variant traditions
in which they have been preserved, as well as the incidence of exegetical
gloss and linguistic assimilation, comprise the areas of investigation under-
taken in the first part of these studies.

1 See below, IV pp. 136-7.
4SS90T6 B

2

QURANIC STUDIES

Four characteristic examples are found in the imagery appropriate to the
theodicy: retribution, sign, exile, and covenant. These themes, which
constitute by far the major part of the Quranic message, depend upon a
limited lexical range with correspondingly high ratios of frequency and
distribution. The result is not unexpected: a very repetitive style which
could indicate either a long period of oral transmission or an original
series of unco-ordinated pericopes, or both. 1 The imagery associated with
divine retribution turns principally upon four substantives : umma (nation),
awwalun (predecessors), qarn (generation), and qctrya (abode), accompanied
by such finite verbs as khala, mada, and halaka (signifying decease). Of the
many contexts containing these locutions those which are not anonymous
were analysed by Horovitz, described as Strqflegenden, and ingeniously
related to the seven mathdni(Q. 15 : 87, possibly 39: z^). 2 Emphasis upon
the dramatis personae of those Straflegenden, which I should prefer to call
retribution pericopes, led Horovitz to postulate a number of Biblical
caiques, in turn adduced as evidence of the Arabian prophet's increasing
knowledge of Hebrew scripture. This kind of argument was a corollary of
that scholar's acceptance of the Noldeke-Schwally chronology of revelation,
a feasible but hardly the only method of interpreting the Quranic data. 3

The phrases cJs. Jj 3ut (Q. 2: 134, 141), J*.f aJI j£) (Q. 7: 34,
10: 49), J^.j 2L,f JiO (Q. 10: 47, cf. 16: 36, with shahid 16: 89, with
nadhir 35 : 24), ^ Jl UL.jf JuJ (Q. 16: 63, cf. 29 : 18, 41 : 25), etc.,

are to be understood as hortatory or admonishing, and of eschatological
rather than historical significance. Of similar application is imagery de-
veloped round the concept awwalun, as OjJjVl Jl* L> <J^-» (Q- 2 3 :
81), c^/i\ iu. o5i (Q. 8: 38, cf. 3: 136), ^Jj^fl J>\J VI iJUb h[
(Q. 6: 25), ^ J4j> UL«j! jiJj (Q. 15: 10, cf. 54: 51),
^jJjVl Ji* Q. 43: 8, cf. 13: 6), in which the implicit reference to history
is attenuated by the eschatological framework. The locution USCUl ^5"
jy ^* (Q. 6: 6, 19: 74, 98, etc.) is always anonymous, as are
almost all constructions based on the plural qurun (e.g. Q. 10: 13, 32: 26.
36: 31, etc.). Similarly UL5CUf hj ^ (Q. 7: 4), i>y & b53u.f Uj
(Q. 15 : 4), and with the plural ^1 (Q. 6: 131, 28: 59).

It may seem useful to distinguish between these general thematic formu-
lations and those narrative conventions employed to introduce or to con-
clude actual accounts of divine retribution found in the Qur'an, such as

Ji U t Usf dLLc ^aSi <i)ijS~ (Q. 20: 99, cf. 7: 101, 11: 100, etc.),

1 See below, pp. 20, 46-52. 2 Untersuchungen, 10-32.

3 See below, pp. 38-43.

REVELATION AND CANON

3

or 'ij^s. (< 4va^ai ,j olf Jui! (Q. 12: in). The phrase jl5" Ij^iuU

J^s £_» ^jJI SJU is either of anonymous application (Q. 30: 42) or

specific (e.g. Moses, in Q. 40: 21), like several of the narrative formulae. 1
Closely related to the imagery of retribution and containing an allusion to
what might be regarded as its Sitz im Leben is Q. 27:34 lit iJ^LJl j) cJli

Ojljiij dJjjJ'j ili! l&l&l ojr&l UjJujt 5j^s m the context

of Solomon's letter to the queen of Saba' a threat of destruction is seen
to be of secular origin, but employing the 'abode' (qarya) image, the one
of most frequent occurrence among the Quranic formulae for divine
retribution. It is, on the other hand, hardly likely, in view of the basmala,
that we are here offered even the reminiscence of an authentic docu-
ment, and the expression 'when they enter a village they destroy it' is
probably to be interpreted as Nachdichtung derived from precisely that
formula. 2 It would anyway appear that the conventional designation of this
theme adopted by exegetes was umam khaliya (the nations which have
perished), e.g. by Muqatil b. Sulayman ad Q. 18: 9 and 55 to gloss dydtind
(our signs) and sunnat al-awwalin (the way of the predecessors), respec-
tively. 3 Umam khaliya may be thought to exhibit the motif: Ubi sunt qui
ante nos in mundo fuere characteristic of Wisdom literature, despite absence
in the Quranic text of an exact equivalent to Ubi sunt.* Identification and
historicity of the umam khaliya were problems posed and solved in the
literature of haggadic exegesis. 5 In Muslim scripture itself they represent
merely perpetuation of a literary type.

The nations perished in many ways, each an expression of divine
vengeance. Here the Quranic vocabulary draws amply upon the lexicon of
literary Arabic. Based upon only a few finite verbs, of which the most
common artja'ala, akhadha, arsala, and waqa'a (all containing the conno-
tation : to set in motion), the primary concepts 'dqiba (destiny) and 'adhdb
(retribution) are realized as a series of variations upon the theme of natural
disaster: e.g. rajfa (Q. 7: 78), tufan (29: 14), rijz (2: 59), matar (26: 173),
hasib (17 : 68), rih 'aqim (5 1 : 41), qasifmin al-rih{\'] : 69), hydra win sijjil(i 1 :
82), hijdra min tin (51 : 33), sayl al-arim (34: 16), rih sarsar (41 : 16), fdra
'l-tannur (n: 40), tnd' munhamir (54: 11), jardd zva-qummal wa-dafddi'
wa-dam (7: 133), sinin wa-naqs min al-thamardt (7 : 130), sSiqa (2: 55); but
occasionally also as more direct intervention: e.g. la'na (28 : 42), sayha (1 1 :
67), zajra^T- 19)> sdkhkha(8o: 33), used to convey the notion of curse, as

1 See below, pp. 18-20.

2 Pace Horovitz, op. cit. 35.

3 Tafsir, MS H. Hiisnii 17, i67 v , i7o v ; see also the references in Goldziher, Abhand-
ungen II, xvi-xviii: quburiyydt.

4 Becker, 'Ubi sunt', esp. 509 ; cf. Horovitz, op. cit. 16, 21, 29 ; and for poetic treatment,
Hirschberg, Lehren, 53-7.

5 See below, IV pp. 135-6.

4

QURANIC STUDIES

well as generic terms like ta'if(68: 19), taghiya (69: 5), batsha (44: 16), or
karb (37: 115). The significance of this imagery is typological and the
punishments exemplary, of the kind referred to in Q. 13. 6 ^ Jij
c/A^aJ i and 34: 1 9 d*j .> U.f UL*»j . Both frequency and distribution

ol the selected lexica suggest originally distinct pericopes.

To be contrasted with the exempla afforded by the umam khaliya are the
concepts of trial and ordeal exhibited in the verbs bald and fatana with God

as subject, e.g. p5oj ^SOi Jj (Q. 2: 49), a^cA U lit jUJ'i/l Lti

<Uj (Q. 89: 15), p^LS ^ ^jjJI £s oiJj (Q. 29: 3), and especially ^Jl Jf

UJIj i^j j-JJIj j^JL (^^jj OjJI SJulS (Q. 21: 35). Similar

constructions with adalla (cause to err) may also be adduced, e.g. tUl-if

t UL> JOJt J^u (Q. 74: 31). Divine justice is here mitigated by what
seems clearly to be a reflex of the Biblical election tradition, evident in
passages where Abraham (Q. 2 : 124) or David (38 : 24) or Solomon (38 : 34)
is object of the trial. Not unexpected is the occasional allusion in these
contexts to its corollary, the remnant tradition (baqiya, baqiya, bdqiyun):
general in Q. 11: 116 j^Jjl ^5CLi ^ djyi\ and specific in 26: 120 and

37: 77 (Noah), 43: 28 (Abraham), 2: 248 (Moses and Aaron), 69: 8 ('Ad),
and 53 : 51 (Thamud), the two last negatively expressed and thus belonging
to the umam khaliya. The appearance in Muslim scripture of these two
Biblical motifs may, in my opinion, be interpreted as Bildungserlebnis
imperfectly assimilated.

Of quite different character is treatment of the retribution theme
reflected in the locution ayydm alldh (days/battles of God). In Quranic
usage yavm may be merely chronological and quantitative, as in Q. 2 : 196,
41 : 9-10, etc., but the context is more often eschatological, e.g. ol Jt» {j*
<UII ^» <U V J| L (Q. 30: 43, cf. 14: 44) and in phrases like ^uj
<S-*y_ e.UJ (Q. 32: 14) and ^ .jLw ^50 (34: 30). Its typical function as
complement to umma is evident in SJUJ| ^b'i/l (Q. 69: 24) and Jl*
j^Li ^ (10: 102). In Q. 45: 14 aM\ j.b! dy?ji ^ oA" th e

plural ayydm is the equivalent of the eschatological singular in the
examples above and of the Biblical nw DT\ With one exception (Job 24: 1)
Biblical usage favours the singular, which may of course refer to a
historical event, e.g. Isaiah 9: 3 f*ia DT", a synonym of fianVa (Judges 8:
13). But the days/battles of Israel are also those of God (Numbers 21 : 14
m?T nan^a) a concept very likely the source of ayydm not only in Q. 45 : 14

(above), but also in 3: 140 ^Ul ^ l$Jj|Jb fb^l dUJj and 14: 5 ^.1
ail I |»Ut ^ J>ij jjJt J! oUIkJI ^ duy, the latter addressed to Moses.
Now, the existence in Arabic of a profane tradition (ayydm al-'arab) might

REVELATION AND CANON 5

be thought to provide an alternative source for the Quranic locution, much
as it appears to have been the source of 'epic' motifs in classical Islamic
historiography. 1 The possibility of contamination may not be excluded,
and conceptual transfer within the Arabic lexicon from secular to divine
(battle) is scarcely difficult. Against the non-occurrence of malhama in
Muslim scripture may be set its primary use (in the plural : malahim) as
designation of apocalyptica, 2 as well as the semantic equivalence of the
cognate roots felt by Saadya in his translation of Psalm 109: 3 and of
Exodus 15 : 3. 3 Like the remnant imagery contained in baqiyaiyeter/she'ar,
the equivalence ayyam.milhamot exhibits theological elaboration of a
Wisdom motif (Ubi sunt), of which a search for the origins might seem
superfluous. In prophetical language, as opposed to that of apocalyptic, the
imagery of retribution belongs to the oracular judgement (e.g. massa in
Isaiah and the later prophets), and it is that literary type which is per-
petuated in the Qur'an.

For expression of the 'sign' as manifestation of the deity and as creden-
tial of the prophet, the lexical range is somewhat wider than for the theme of
retribution. One may distinguish between Quranic terms for sign and
descriptions of them. The latter fall into four groups, of which the most
frequent is imagery drawn from the phenomena of divine creation, ex-

ft * *"

hibiting variations upon the motif of Q. 3: 190 j^Vlj oIja*J! j jl
^UVl Jj^ oUV jL^JIj JJI o^A^i-lj. In addition to the blessings and
disasters of nature events of an unusual character (khdriq Kl-'ddd) are
introduced, e.g. splitting of the moon (Q. 54: 1, cf. 55: 37, 84: 1) and
crumbling of mountains (Q. 77: 10, cf. 59: 21, 7: 143). Different from but
related to those are instances of more personal intervention in the laws of
nature by its creator, e.g. nourishment (Q. 7: 32, cf. 2: 257), clothing
(7 : 26), treasure (1 1 : 12, cf. 9 : 34, 43 : 33-5), hosts (36 : 28, 48 : 4, cf. 3:12,
5 : 52), sanctuaries (2 : 158), but also the exempla of salvation history: Noah
and his people (25: 37), Joseph and his brothers (12: 7), Jesus and his
mother (23 : 50), the men of the Cave (18: 9).

Rather more important to the elaboration of dogma are two other
categories of sign, which may be designated respectively guidance and
scripture. To the former belong ruh, amr, kalimajhalimdt, hukmjhikma,
shir' a, minhdj, misbdhlnUr, maw'iza, wa'djwaid, bora a, adhdn, but also
rasuljrisdla, bashir, nadhir, nabi/nubuzowa, and malakjmala^ika. To the
latter may be reckoned kitab/kutub, qwfdn, injtl, tawrdt, furqdn, zaburj

1 See Caskel, 'Aijam al-'Arab', 1-99; and below, IV pp. 125. 140.

* Cf. Dozy, Supplement ii, 522, and kutub malhama with sefer milhamot, Numbers 21 :
14; and Goldziher, Richtungen, 57, but also 66 n. 5, where, however, I should equate
malahim with the Quranic ayyam alldh, not merely with 'Nachrichten apokalyptischer
Art'.

3 Lauterbach, Saadja, vi, 25 n. 3; Rabin, Qumran, 1 18-19.

6

QURANIC STUDIES

zubur, suhuf, sura/suzcar. It would be misleading to propose a rigorously
schematic distinction between the two sets of lexica, in which semantic
boundaries tend to be blurred by contextual similarity. References in
scripture to scripture pose a particular kind of problem, related to the role
of proof-texts in Islamic prophetology. 1 Collocation of sign as guidance
and sign as scripture is of comparatively high frequency in the Qur'an, e.g.
kalimat with Utah (Q. 18: 27), hikma with kitab, tawrat, and injil (3: 48,
5:110, etc.), hukm with kitab and nubuwzoa (3 : 79, 6 : 89, etc.). Ruh and amr,
though not formally linked with 'scriptural' signs, appear in contexts
determined by finite verbs signifying revelation, e.g. awha (Q. 42: 52) and
anzala (65: 5), as do general terms for 'guidance' like ruskd, huda, bayan\
tibyan, and even rahma (mercy). Rasul (messenger) is of course found in
combination with nearly all the words of both categories. From collocations
of this kind one might infer that the Quranic concept of theophany is
basically scriptural, an argument commonly referred to a tradition of
rhetorical accomplishment among speakers of Arabic. 2

'Sign' is generally conveyed in Muslim scripture by aya, which in Q.
38 : 29 may be understood to signify 'verse of therevealed book' oUJjjl t—ilxS"
AjbT Ij^J^J «±JLJ1. Elsewhere, however, aya is merely 'exemplum'

and from analogous contexts attracts a number of synonyms, e.g. 'alamat
(Q. 16: 6, curiously only once, though this is the most common equivalent
in exegetical literature), 'ibra (12: m), uswa (60: 4), hadith (45: 6), mathal
(43: 57), tadhkira (20: 3), dhikrfo: 58), dhikra($o: 37), bayyina (20: 133),
burhdn (4: 174), sulfdn (30: 35), naba (38: 67), sha'd'ir (22: 36), ashrat (47:
18), dthar (30: 50). In Q. 48: 29 sima and athar are not synonyms of
aya, but signify rather 'imprint'. 3 Save for Surat Rahman (verse 13 and

passim: L5Cj e.sl ^Li) it is ayajayat which bears the burden of a

Quranic refrain: SjV dlli> J jl ( sm g- Q- 2: 248, siiras 16 and 26 passim;
plural 10: 67, sura 30 passim; six occurrences with 'ibra, e.g. 3: 13). The
locution belongs to the imagery of prophetical expression, e.g. Isaiah 37: 30
msn -]b nn, and may announce a miracle: e.g. Isaiah 7: 14 Xin vtH p"?

niK 00b and Q. 18: 9 bJLT ^ 1^15" pJj\j vJ^53l v 1 **^ ol ft

It is precisely in contexts characterized by the equivalence sign: miracle,
demanded as credential of the prophet, that the lexical range is widest.
Introduced by such conjunctive particles as lawld, lawma, law anna, hattd,

' See below, II pp. 63-5.

2 See below, II pp. 79-81.

3 Pace Hirschfeld, Researches, 96-7 n. 75 where read sima or aya, but see also 45-6,
60-1 ; cf. 'urtwan al-sujud in Noldeke, Delectus, 77, line 14.

+ Cf. Westermann, Grundformen, 113-4; Quranic usage, it may be observed, partakes
of both Heilsankundigung and Gerichtsankiindigung.

REVELATION AND CANON

7

mata, or by a simple imperative, the demands include gardens, springs,
fire, victory, angels, messengers, and (!) scripture (kitab, suhuf, sura,
qur'dri). The formulae range from a modest reproach, e.g. juL& Jj^'f VjJ
Ajj ^ 5jT(Q. io: 20, 13: 7, etc., cf. 20: 33) to the elaborate inventory of
Q. 17:90-3:

We will not believe you until you cause a spring to erupt from the earth, or till
you have a garden with palms and grapes in which you can make rivers flow, or
till you cause heaven to collapse, as you claim, or bring us face to face with God
and the angels, or till you own a house of gold, or ascend into heaven, nor will
we believe your ascension until you send down to us a book that we can read.

The scripture: miracle imagery is of course that underlying the tahaddi
(challenge) verses. 1 Relation to the Quranic version of the Ubi sunt motif
was achieved by allusion to earlier scriptural ( !) revelations, e.g. Q. 21 : 5

jjJjVl J—jf Ui" XjU LxjLJJ j£.\Jl ja Jj el j&\ Jj jO^I djUyif ]ji\S Jj and
37: 168 jjJj Si ^* I jS^i L'JUft o! ji. In this context the not quite un-
expected disclaimer of miracles and refusal 'to grant a sign' (cf . John 6 :
28-32, Matthew 16: 1-4, Luke 11: 29-32), e.g. Q. 7: 203 lilj

I VjJ IjJU AjIj and 40: 78 AJUl oil* ^' j\i jl d& L>j

is further evidence of conformity to scriptural archetype.

A third theme contributory to the Quranic theodicy is that of exile. A
primary source for the imagery is a description of Abraham's withdrawal in

search of God from the environment familiar to him cJl l_-jMjI Jli

f»*3j^b U-« Or^b tih^j^ axj ^ b

jji Cy* OjiJo Lj (Q. 19: 46, 48). Use of i'tazala (withdrew) may be

compared with Q. 18: 16, where the flight of the men of the Cave from
persecution is so named. But it is the verb hajara, imperative here (leave
me!), which exemplifies the Quranic concept of displacement, whether
for the sake of worship, of redemption, or of bearing witness (martyrdom).
Employment of the imperative for that classical instance of exile (cf.
Genesis 12: 1 lekh lekha) becomes in Q. 73: 10 an expression of general

paraenesis ^ 1 j**> j** \j Ojiyi U ^s. (cf. lekh/halokh,

Isaiah 6: 9, Jeremiah 2: 1, Ezekiel 3: 4, Amos 7: 17, etc.). The active
participle of the first form appears in Q. 59: 9, but of most frequent
occurrence are the perfect, imperfect, and active participle of the verb's

third derived form, e.g. Q. 2: 218 IjjjtUj ji^b (ji-^l ol

■UJl J and 4: 100 I jjS' Lc-Ij-* ii tJsr*" <i Cr*J

1 See below, II pp. 78-9.

8

QURANIC STUDIES

aJj-^jj «U)) Jl Ij»U-* a^j (j-. ^ej^yj iu.j. The parallelism of yuhajir
and yakhruj exhibited in Q. 4: 100 is attested elsewhere, e.g. 2: 74, 60: 1,
and in 4 : 66 with the imperative ukhruju. A fourth derived form with God

as subject appears in Q. 8: 5 <j?tJb di^j ^ <iljj t^j^f l*^- The latter
construction, accompanied by the participle muhdjir, is especially con-
spicuous in a context containing the only Quranic instance of ]ala as exile

UjOI J OkJI p^ift -*JU 1 u_-=5" jl ^jJj (Q- 59 : 3» cf. verses 2,

8, 9, 11). There, the notion of exile, though an act of God, is not an
imperative addressed to a prophet, but belongs instead to the vocabulary
of prophetical threats (cf. golah, Nahum v: io, Amos 1 : 15, etc.). Surat
Bard' a contains a number of parallel constructions with kharaja and nafara
signifying movement towards, or at the behest of, God, e.g. the imperative

in Q. 9: 38 JOJl J™^, J Ij^l ^50 JJ 151 (cf. verses 39, 41-2, 46-7, 81,

122). The imperative of farra in Q. 51: 50 <OJ| J I is isolated. Of

Moses and Lot the verb asrd is used, often glossed 'in the night', e.g. Q.

44: 23 ^KJ iS^^i (cf- 30 : 77> 2 6 : S2, and lekh lekha in Exodus 3: 10,

16, 4: 12, etc.), and n: 81 Jail ^ ^kk d^t j^i (cf. 15: 65). It is that

locution which was employed for the isrS verse (Q 17: 1), a Mosaic
reference appropriated in exegetical literature to the biography of the
Arabian prophet. 1 The divine imperative may also express the notion of
movement combined with the imagery of retribution, as in Q. 3: 137

^JiSCJ! XJlt q15" IjjkjU <j*j"^l <j (cf. 6: 11, 16: 36, etc., and

in the imperfect : 12 : 109, 30 : 9, and passim), and especially 34 : 18, following
upon the devastation wrought by the deluge (verse 16: sayl al-'arim). The
verb sara occurs once (Q. 28 : 29 Moses) in the perfect, and to indicate
movement at the command of God, a notion symbolized in exegetical
literature by the non-Quranic substantive hijra. 7 -

The unilateral character of the imperative is reflected in the fourth
theme here proposed to illustrate the Quranic theodicy: the covenant. The

locus classicus is Q. 3: 81 /^--Jl cjl£~i «U)1 J^f ilj amplified in 33: 7 by

reference to Noah, Abraham, Moses, and Jesus, and employed in the
exegetical tradition to depict the origins of prophecy in Arabia. 3 The finite
verb form akhadha (took), with God as agent or one of His retributive
'signs' as source of action (e.g. Q. 11: 102, 22: 48, and 11: 57, 2: 55,
respectively), invariably expresses unilateral imposition, and is thus used
in the two passages cited above, as well as in most other locutions based on
mithdq, when the latter refers to a covenant between God and man

1 See below, II pp. 67-8.
3 Ibn Hisham, Sira i, 233-4.,

2 Cf. Wensinck, 'Propheten', 180-90.

REVELATION AND CANON

9

(Q. 2: 63, 83, 84, 93, 3 : 187, 4: 154, 5 : 12, 14, 70, 7: 169, 57: 8). The divine
covenant is also called 'ahd, for which the perfect of the verb 'ahida, with
God as agent, expresses a unilateral function, e.g. Q. 2: 125, 3: 183, etc.
(36 : 60 has lam with the jussive). A complementary and bilateral concept of
'ahd is exhibited in the use of third form perfect 'ahada, with God as
object, e.g. Q. 9: 75 -UJl JjoU ^* ftg^j- While the substantive 'ahd seldom
occurs with the finite verb 'ahida (but cf. Q. 2: 124-5), there is a single
instance of mithaq governed by an otherwise unattested wdthaqa, with God
as agent (Q. 5 : 7). As designations of the divine covenant 'ahd and mithaq
can be synonymous, e.g. in Q. 2: 27 as\J±, o»j ^ «U)I Qj^>&> ^jJI

(cf. 13: 20, 25), though it is not impossible that here mithaq refers to
(written ?) corroboration of a covenant already articulated, an interpretation

supported to some extent by Q. 7: 169 i_>US3l (Jl2^ ^Is. iijj f, as

well as by derivatives like wathaq (47 : 4, 89: 26) and ('urwa) wuthqa (2 : 256,
31 : 22). In his translation of the Psalter, Saadya rendered 111:5 B^iS 1 ? *13F
WIS in Arabic as mn» T3l6N "-Vs M JTOIV Like Biblical ferd, both
'aArf and mithaq may designate a covenantal relation between men, as in
Q. 8: 56 jyfcj V pAj 'ij* JT J j>»aAL> f»J f»4^» oJaU ^jjJI

and 8: 72 ^4^5 ^£,0 j.y "ill ^aJI j^Cl*s ^jJI J pi'^a^J jlj

(JjlS^, respectively. In the former passage use of the third derived form
('dhadta) is characteristic of the secular application of 'ahd (e.g. Q. 2: 100,
177, 9: 1, 4, 7), as it is also of the bilateral covenant (cf. Q. 9: 75 above,
and 16: 91, 33: 15, 23, 48: 10) between God and man. For the latter an

eighth derived form of akhadha is also attested, e.g. Q. 19: 78 i_^~jUI ^U»f
iJifrc i>*vJI j^l fl (cf. 19: 87, 2: 80).

The Quranic covenant imagery may reflect development from profane
legal terminology by the introduction of divine sanction, though it is of
interest to note that in the 'Umma document(s)' preserved in the Sir a of
Ibn Hisham neither 'ahd nor mithaq occurs : the finite verb 'ahada ( !)
appears once and the document itself is called sahifa, a neutral term. 2 On
the other hand, conceptual transfer from profane to divine imagery, and

vice versa, may be inferred from modification of a formula such as Jl$*JI jl
Vj-vw* olf (Q- 17: 34) to AAJl J^a ofcj (33 : 15), or from the inter-

polation min allah after each of the only three instances of mawthiq (Q. 12:
66, 80), the last of which, significantly, is object of the verb akhadha with
Jacob as agent. The observation of Horovitz on a distinction between

1 Lauterbach, Saadja, x, 34 n. 4.

2 See Serjeant, 'Constitution', 3-16; it may well be that the Quranic plural fuftuf is
a reference not merely to 'scripture' but to covenant, cf. esp. Q. 87: 18-19, i- e - 'the first
suhuf, those of Abraham and Moses'.

IO

QURANIC STUDIES

mithaq as caique of berit and/or diatheke and mithaq as 'Terminus der
arabischen Rechtssprache' is superfluous: berit readily provides a Vorlage
for either usage, and loci probantes for secular treaty relations will not be
older than the Qur'an document. 1 Similarly Jeffery, in a rather impression-
istic treatment of some of the covenant verses, asserted that mithaq and
'ahd were also employed outside scripture (but when ?). The description
there of accompanying finite verb-forms is inadequately differentiated, and
to adduce Q. 6: 89 (3 : 81 would have been more appropriate !) as belonging
to the covenant imagery must mean that 'covenant' included the gifts of
scripture and prophecy, and will thus have been relevant only to Jews and
Christians, not to all of the umam khdliya. 2 Elsewhere the author's search
for parallels to/sources for the Muslim concept of scripture took him far
beyond the Judaeo-Christian tradition. 3 Such investigation may be of
considerable value for studies in comparative religion, 4 but of rather less
for the-lkerary analysis of scripture, especially in the light of the very
explicit Quranic data. The source of the covenant imagery was clearly
Biblical, and predominantly Pentateuchal. That the Mosaic covenant
reflected in Muslim scripture should exhibit haggadic accretion and even
conceptual deformation is hardly surprising, but worthy of remark:
neither the Deuteronomic nor the Rabbinic interpretation of covenant was
restricted to the express content of the Decalogue. 5

Some evidence for determination of a Sitz im Leben may, however, be
available for other Quranic terms appropriate to the covenant relationship.
As against two instances, both secular, of dhimma (Q. 9 : 8, 10), the Umma

document contains oJbxIj <U)I iL,S jlj !and the locution habl allah (Q. 3:
103) is paralleled in 3: 112 by habl min allah and habl min al-nas, with
which may be compared hibdl in the account of 'Aqaba. 6 Moreover,
amana, used of the covenant between God and man in Q. 23 : 8 and 70: 32
Oj*lj j^aA^cj p^i'ULi t>>jJlj ( c f- 2: 283 and 4: 58, where it is
exclusively secular; but also Nehemiah 10: 1, in which the purpose of
ruSX was adherence to God's law), is widely attested in later chancery
usage. 7 Other essentially neutral terms, like bdya'a (acknowledge authority)
and ayman (oaths), may in Quranic usage be lent divine sanction by

occurring in context with 'ahdzni its derivatives, e.g. Q. 48: 10 ^ jJI j]
<U)I «uU -AaU Lj Jjf ^.j . . . OjajLj U! dijjjuLj and 16: 91
U^Tji' 0^>jI pJOAU lil JUDl J^jo, respectively.

1 Untersuchungen, 51.

2 'Scripture', 119— 21, 127-9.

3 Jeffery, op. cit., 41-55, but cf. also 202 n. 40.

4 See below, II p. 61.

5 Cf. Obermann, 'Agada', esp. 29-47.

6 Ibn Hisham, Sira I, 502, 442, respectively.

7 Qalqashandj, §ubh XIII, 321-51.

REVELATION AND CANON

1 1

Covenant is related in the Quranic theodicy to the exempla of salvation
history, to those representatives of the umam khaliya from which a rem-
nant survived. Reference is general in Q. 33: 15 ^ jUII Ijjublc

«U)I ofefj J I 6>JjJ V Jc> (cf- 2: 286, in which covenant is
depicted as 'burden'/'yoke': isr), and specific in 33: 7 ^.jJI ^ ilj

The context is didactic rather than historical, and the concept limited
to the Patriarchal and Sinai traditions. Allusion to Adam's negligence

lijz <U -XpbJ |^Jj JJ ^ |<oI (Jl Ua^* JiJj (Q. 20: 115) reflects an

admonition attested elsewhere in prophetical literature, e.g. Jeremiah 22: 9
and Hosea 6:7.' Exhortation to observe and to keep the covenant (ra'd/
awfa) is frequent, but it is the Patriarchal and Mosaic covenants which
were to be observed. A single reference to a Christian covenant (Q. 5 : 14)
like inclusion of Jesus in Q. 33 : 7 (above), represents chronological exten-
sion, not historical development. The complementary concepts of fulfil-
ment and renewal characteristic of the writing prophets (Jeremiah 31:31
HEnri rVia) and of the evangelists (Luke 22: 20 wuvq Sia&j/oj) are not
stressed in the covenantal imagery of the Qur'an. 2 Unlike the conceptual
distinction recognized in Old Testament theology between the Deutero-
nomic and Priestly notions of covenant, 3 the difference of semantic
functions between Quranic akhadha and 'ahida, on the one hand, and
'ahada on the other, is almost purely formal. Possibly separate traditions
of bilateral and unilateral covenants coalesce in the Quranic imagery
to produce the concept of submission explicit in the term islam. The
covenant terminology might appear, moreover, to support the derivation
of aslama from saldtn proposed by Lidzbarski, namely, 'to enter into a state
of peace', or perhaps 'of salvation', exhibiting the equivalence berit:
shalom* Reflex of berit is exhibited in two occurrences of the cognate

bora a: Q. 9: 1 l yS' J ^ui\ ^* ^oaU ^jjJI ^1 aJ_^j <U)| ^ and

54: 43 J ig.\ji ^1 p£iJjl ,y j^S. pS'jiiS'], the former in a

context (Q. 9: 1-10) which provided the locus probans for juridical theory
on the scope of secular treaty relations. The clearly formulaic expression
of the entire passage might, indeed, be thought to provide some evidence
of a Sitz im Leben for the covenant as set out in Muslim scripture, similar

1 See Speyer, Erzdhlungen, 66-7.

2 Cf. von Rad, Theology ii, 212-17, 247, 268-72, but also i, 306-18.

3 Von Rad, op. cit. i, 129-35.

* 'Salam', 85-90; Pedersen, Israel i-ii, 263-335; and Kunstlinger, 'Islam', 128-37,
esp. 133-4; rejected by Horovitz, Untersuchungen, 54-5, and Katsch, Judaism, 104-5,
as well as by Bravmann, who detected a reflex of Arabian conduct in warfare, 'Background',
324-43-

12

QURANIC STUDIES

to the role of legalistic and cultic elements in the formulation of the Hebrew
covenant tradition. 1

• • • • •

The presence in the Qur'an of literary convention was recognized by
exponents of what I have described as rhetorical exegesis. 2 The schemata of
revelation include a number of conventions typically employed to introduce
the major themes of the Quranic theodicy. Exhibiting a comparatively
limited lexical range, those formulae serve to confirm the impression of a
composition made up of originally unrelated pericopes. From the very
high ratios of frequency and distribution e.g. for qui and ayyuha, one must
expect a degree of contextual overlapping. From a formal point of view,
however, the conventions can be identified as apodictic, supplicatory, and
narrative. As such they may be regarded as symbolic of the basic, often
minimal, units of expression in prophetical literature. Seldom reducible,
they frequently mirror a pre-literary form, and occasionally a discernible
function in language other than scripture. For example, the covenant with
Adam (Q. 20: 115) referred to above was described by Zamakhshari as

reflex of a royal messenger formula • [ frAl>U>jj iJ^LJI jJjf J Jll>
jUI j^j <uU ^yj aJI yjj Jl dM- 3 Similarly, Q. 27: 34
depicting the arrival of Solomon's letter in Saba' might be thought to
contain incipient elements of the Quranic retribution imagery. The
problem here, and in general, is the difficulty in dating loci probantes from
profane Arabic literature. In its earliest stages that literature is essentially
exegetical, even when not avowedly concerned to elucidate scripture, and
genuinely independent witness to the usus loquendi is indeed rare.

Easily the most expressive example of the Quranic messenger formula is
Q. 19 : 17-21, in which Mary is informed by the Spirit of God (traditionally
understood to be the angel Gabriel) that she is to bear a son. The passage
may be analysed as follows: 4

v. 17 Commission (Beauftragung)
v. 19 Delivery (Vberbringung)

v. 21 Message (Ausrichtung)

The utterance of the messenger (Botenspruch) is contained in verses 19

1 Cf. Buhl, 'Kuranexegese', 97-106; and Koch, Growth, 21-a, 29-33.

2 See below, IV pp. 227-39. 3 Kashshaf in, 90-1 ad loc.
4 Cf. Westermann, Grundformen, 72-3 .

REVELATION AND CANON

13

and 21 (verses 18 and 20 report Mary's reaction to the incident), and the
messenger formula proper (Botenformel) may be isolated as dbj Jli tU)
i.e. mrp fD. 1 The locution kadhalika qala rabbuki{kd) is employed also
in Q. 19: 9 for a similar message delivered to Zechariah, and in 51 : 30 to
Abraham, suggesting thus a stereotype for identical situations. Of related
application is -UJl o-A5" cJa- tUI^S" in Q. 10: 33 and 40: 6. In what maybe
the same function kadhalika appears in a pausal context (followed by
wa-qad) in Q. 18: 91, and as corroborate in the queen of Saba"s obser-
vation on the conduct of kings: 27: 34 j^Uij dJUJiTj. 2 The demonstrative
kadhalika, often synonymous with kama in reference to an earlier,
analogous event, may occur with finite verbs and God as agent, e.g. Q. 1 1 :

102 (.syJI jsj lil dlllTj and 18: 19 dlloTj, with which

may be compared 2: 286 UL» ^ (ji-jJI L5" and 4: 47 £«J LjT

c~*JJ <_ibey>!. More often, however, the function of kadhalika is that of
presentative, employed to introduce utterances both exemplary and general,

e.g. djU -Oil ^ dJUlT(Q. 2: 187, 3: 103, etc.), oL^I J-^ii dU-iTj
(6: 55, 7: 32, etc., and with nusarrif 7 : 58), or the refrain iHJjd Ul

( ^L t ^ E J| (37: 80, 105, 110,121, 13 1). 3 As in the prophetical literature of the
Old Testament the messenger formula, also designated formula of legiti-
mation (corroboratio), may be merely understood, and constructions in-
cluding only the term of commission, e.g. laqad atayna, laqad arsalna
(Q. 2: 87, 27: 15, 31: 12, and 30: 47, 40: 23, 29: 14, respectively), are
common in Muslim scripture.

The most frequent by far of Quranic apodictic formulae is qui (say,
speak !), a locution which dispenses altogether with the messenger formula
as set out above. The observation of Horovitz that all of the Qur'an must be
regarded as the utterance of God is doctrinally sound but for a form-
critical approach very frustrating. 4 Since, however, it has never been
possible to insist that production of the Quranic revelation was analogous
to that of the Hebrew prophetical literature, problems pertaining to the
development of literary forms are difficult to define, let alone solve.
Westermann's chronological formulation 'prophetisches Wort als Boten-
wort' presupposes just such a discernible development, attested by the
various forms in which the word of God was uttered, from the Pentateuch

1 Westermann, op. cit. 107 ff.

2 According to Ibn Qutayba, wa-kadhdlika in Q. 37: 34 was preceded by a pause and
introduced a fresh proposition, uttered by God, Ta'viil 226-7. Similarly, a pause after
kadhalika in the Botenformel, e.g. Q. 19: 9, was the subject of exegetical dispute, see
below, IV pp. 224-5.

3 Cf. Koch, GrOfvth, 212: lakhen; kadhalika so employed was usually interpreted
hdkadhd, see below, IV pp. 129-30. 4 Untersuchttngen, 5.

QURANIC STUDIES

through the prophetical books to Chronicles. 1 By contrast, the Qur'an
exhibits a variety of recognizable literary forms in no recognizable order.
Less dogmatic than Horovitz, Suyiitl adduced five passages in Muslim
scripture whose attribution to God was at least disputed c Q. 6: 104, 114
were the words of the Arabian prophet ; 19 : 64 (but, curiously, not 19:9,21
and 51: 30) were the words of Gabriel; 37: 164-66 were ascribed to the
angels; finally, verse 4 of the Fatiha may have been uttered by the faithful
('ibad) or could, by insertion (taqdir) of the imperative qulu, be attributed to
God, a process applicable also to the two passages ascribed to the prophet. 1
The mechanism of insertion, developed by the masoretes, 3 must surely
account for some at least of the 350-odd occurrences of an imperative form
of qala (speak). 4

Its functions are several, but it would be an exaggeration to describe its
over-all effect as successful replacement of typically prophetical expressions
by the direct speech of God. The theological implications of the inverse
development in Hebrew scripture could provoke an impression that
Muslim scripture represented a conscious return to the unmediated
theophany of the Hexateuch. 5 That would, I think, be misleading, even
were the mechanism of divine imperatives consistently applied, which
clearly it was not. The role of the prophet in Islam is too central to allow of
such an interpretation, however much the Muslim doctrine of scripture as
the word of God might appear to lend it support. 6 What might be con-
strued as a logical contradiction between the two views is reflected in the
functions of qui. It is, for example, employed to introduce mention of God
in the third person, as in Q. 2: 120, 140, 7: 28-9, 13: 36, 27: 93, etc.,
though such passages are far fewer in number than those without an intro-
ductory qui, perhaps best exemplified in utterances of the type: wa{kana)
allah 'alimjrahimjhakitn, etc. Qui may preface an apodosis after statements
beginning yas'alunaka (they ask you), often of halakhic content, as in the
series Q. 2: 218-22, or 4: 176 and 5: 4, but also in matters eschatological
(7 : 187) and anecdotal (18 : 83 ). 7 Qui commonly serves to indicate liturgical
instructions, frequently prayer e.g. Q. 3: 26, 10: 104, 13: 16, and especially
suras 1 12, 113, and 1 14. 8 Finally, but most significantly, qui may introduce
statements not predicable of the deity and usually containing finite verb-
forms like 'I fear' (Q. 6: 15), T have been ordered' (13 : 36), *If I err' (34:
50), but also descriptions of the type 'I am only a warner' (38: 65, etc.), as

1 Grundformen, 70-1 (italics mine).

* Itqan, naw' 10: I, 99-101.
3 See below, IV pp. 219-24.

* Cf. Noldeke, 'Zur Sprache des Korans', NBSS, 8; GdQ ii, 36, 42.

5 See Koch, Growth, 216 ff; von Rad, Theology ii, 33-49.

6 See below, p. 38, and II, pp. 55-6, 77-8.

7 For the manner in which yas'alunaka may generate an aetiological legend, see Horo-
vitz, Untersuchungen, 6, and below, IV pp. 122-6.

8 Pace Noldeke, NBSS, 8.

REVELATION AND CANON

15

well as stereotype formulae like 'God suffices as witness between us' (29:
52). It is for these last categories that the possibility of interpolation sug-
gests itself, but to insist upon such in a text characterized by the absence of
logical structure, as contrasted with stylistic homogeneity, would be
superfluous. 1

A somewhat less common apodictic formula is the vocative yd ayyuhd.
Though from the observations so far made this locution could be seen to
require a preceding qui, it is in fact so provided only six times, four of
which contain declarations in the first person that could not be ascribed to
God(Q-7: 158, 10: 104, 22:49, 109: 1-2). It would thus not be impossible
to see in that phrase the primary form of prophetical announcement in
Muslim scripture. Most often the formula introduces an imperative (e.g.
Q- 35 : 3) or a prohibitive (24: 27), but also a conditional construction (8:
29), and occasionally a nominal (33: 50) or interrogative sentence (61 : 10).
The addressees may be either 'you people' without qualification (al-nds, cf.
also al=mala', al-insdn), or 'you believers/disbelievers/Jews/scriptuaries'
(alladhina dmanu/kafarulhadu/utu 'l-kitdb), or finally, one of several
epithets referring to God's messengers, e.g. rasul, rusul, mursalUn, nabi,
muzammil, muddaththir, etc. Utterances addressed to members of the
latter group might of course be interpreted as containing implicitly the
divine imperative, and hence equivalent to expressions prefaced by qui.
Though it is possible to argue that use of the vocative dispenses with a
specific phrase of legitimation (corroboratio), such as the messenger formula,
the two elements are by no means mutually exclusive. 2 Some vocative
constructions appear invariably in contexts containing also the imperative
qui, e.g. four of the five occurrences of yd bant ddam (Q. 7: 26, 27, 31, 35 ;
the fifth is 36: 60). Despite the plural and the assertion of Noldeke-
Schwally to the contrary, I am inclined to relate this locution to the voca-
tive ben adorn characteristic of Ezekiel (2: 1 and passim, cf. also Daniel
8: 17), where modification of the messenger formula is not infrequent, e.g.
3: 11 ro?r tix n»s ro nirVK masi nrcbx man and u: 5 i!3*n
mm

A third vocative found in the Qur'an is the Wisdom formula yd bunayya
(Q. 11: 42, 12: 5, 31: 13, 16, 17, 37: 102), attested also in the plural yd
baniyya(Q. 12: 67, 87 where Jacob addresses his sons; 2: 132 where Isaac
and Jacob are addressed by Abraham). Unlike the corresponding Hebrew
locutions (Proverbs passim: D^a WlSl/ma 157»E?/' , 33 !?»») the Quranic ones

1 See below, pp. 46-7.

" Cf. Koch, Growth, zio-11.

3 GdQ i, 242 n. 1 : their proposed alternative appears only twice in the vocative,
Q. 82: 6, 84: 6 yd insdn, and can hardly be said to represent a functional counterpart
of the common address in Ezekiel; nor, it may be added, was ben adorn in Ezekiel rendered
in the Targum by bar nashajbar nash (i.e. insdn), but by bar adam: see Vermes, 'Jewish
Aramaic', 328.

16 QURANIC STUDIES

are only formally linked to Wisdom: in Q. 31 : 13-13 S*£>JI ^*2J Lj'T Ji23j

•Uj'il ^aaJ (Jls ilj . . . ., and introduce
exhortations and admonitions of a distinctly monotheistic flavour. 1
Similarly related to Wisdom literature is the formula: am hasibajhasibtaj
hasibtum an (e.g. Q. 29: 4, 45 : 21, 47: 29, 18: 9, 2: 214, 3 : 142, 9: 16, and
the variations in 23 : 115, 29: 2 'did you suppose that', etc.), in consequence
of which the prophetical announcement becomes a rhetorical question. 2
This attenuated form of imperative is also encountered in the later prophets,
e.g. Ezekiel 8: 6 nm nxm ntH~p and 8: 17 DTK-p mnn, where apodictic
formulae are often of a character almost conversational. To relate the
juxtaposition in the Qur'an of such a variety of forms to a history of the
prophetical office would be a hazardous, if not quite futile, undertaking.
The conceptual and formal transfer of terminology, very likely in both
directions, between the roles of envoy as royal agent and as divine mediator
had been effected in Semitic literature long before the composition of
Muslim scripture. 3 It is to that tradition, rather than to one fabricated
from the data of Arabic literature, that a literary history of the Qur'an
must inevitably lead.

Analysis of the Quranic formulae of supplication reveals a similar set of
problems. Evidence of such pre-Islamic liturgical practices as might be
presumed to offer a recognizable source for the Quranic references is
encased in a pseudo-historical projection of Islamic scriptural exegesis. Its
classical formulation is found in the Kitab al-Asnam of Hisham b. al-Kalbi
(d. 204/819).+ There Arabian litholatry is described as the consequence of
strife among the descendants of Ishmael in Mecca, their dispersion from
that sanctuary with relics of the Ka'ba in their possession, and continued
observance of the traditional cultic procedures at these several monuments
'in memory of and abiding affection for Mecca'. 5 It is, then, not surprising
to find that such ritual phenomena as honouring the sanctuary (ta'zitn al-
bayt) by circumambulation (tawaf, dawar), station {wuquf), sacrifice (ihdd'
al-budn), and pilgrimage (hajj, 'umra) were in fact derived from the coven-
ant ('ahd) imposed by God upon Abraham and Ishmael at the Ka'ba. Some
admixture of alien and distorted observance (juu ^^J U ^Uol £.)
was inevitable, though such appears to have consisted largely in neglect
of the originally symbolic function of the bits of masonry detached by
departing factions from the sanctuary at Mecca. Originally unitarian
formulae of supplication, like the tasmiya and talbiya, were thus preserved,

1 Cf. Lindblom, 'Wisdom', 193-204 for the link between Wisdom and monotheism;
Goldziher, 'Polemik', 354-5; Horovitz, Untersuckungen, 132-6; cf. also Noldeke, Delectus,
1 line 8.

* Cf. Proverbs 20: 9; and von Rad, Theology ii, 68 on Job 4: 17.
3 See Westermann, Grundformen, 82.

♦ GAS i, 268-71. s Kitab al-Asnam, 6.

REVELATION AND CANON

17

but at the cost of adulteration by polytheistic foci. 1 The strongly exegetical
character of Hisham's description of pagan practices is clear from his
account of the introduction of such by 'Arm%b. Rabi'a/'Amr b. Luhayy,
where the classification of sacrificial animals reflects the allusions of
Q- 5 : J S3 (bahira, saiba, wasila, harm). 1

The use of this and related material by Wellhausen and Lammens
resulted in a portrait of the Jahiliyya that, in accordance with the best
principles of Biblical literary (documentary) criticism by which the earliest
ascertainable form of report becomes synonymous with event, made the
transition from pagan to Islamic worship a logical and almost predictable
process. 3 The antonomastic development which produced Allah as the
Islamic name for God was seen to provide reasonable support for the
social evolution inferred from the exegetical literature. 4 The Quranic
forms of address employed for supplication admit of the same inter-
pretation, and it is thus not difficult to postulate as Sitz im Leben for the
expressions yd rabband (address) and yd 'ibadi (response) the usus loquendi
of the Arabian kdhin* Indeed, the most frequent supplicatory formula in
the Qur'an is rabbijrabband, without die vocative particle but often made
asseverative by a prefixed qui (Q. 13: 30) or qui inna (34: 39). Both rabbi
and rabbana may introduce a petition formulated as imperative, e.g. Q. 3 :
194, 7: 126, 14: 41, 19: 6, 10. The response, introduced by yd 'ibadi, may
also include an imperative, e.g. Q. 29: 56, 39: 10, 53, 43 : 68. One is, more-
over, not surprised to find the locution alldhumma, an alleged vocative
form of allah and attested in the Qur'an (3: 26, 5 : 114, 8: 32, 10: 10, 39:46),
contained in the polytheistic talbiya of Nizar. 6 For the Muslim talbiya,
which is non-Quranic, a secular origin is plausible: liLD is, however,
functionally comparable to Biblical "p*WK (Deuteronomy 33 : 29, etc.). 7

In Quranic usage both rabb and allah (or a pronominal substitute) in
combination with the verbal nouns hamd or subhdn (denoting gratitude or
praise) generate a number of exclamatory constructions, e.g. al-hamdu
lillah, lahu 'l-kamd, bi-hamdi rabbika, bi-hamdika, etc., or subhdna 'lldh,
subhdna 'lladhi, subhdna rabbika, subhdnaka, etc. All of these reflect an
obviously liturgical origin, though hardly attested in Hisham's description
of Arabian paganism, save in the antithesis kufrdnika -.subhdnika and the
parallelism tasbih.tahlil, both in verse. 8 With the exception of Q. 3: 188
(and possibly 61:6) finite verb forms of the root h-m-d do not occur in the
Qur'an, though of the root s-b-h finite forms are common. An interesting

' Kitdb al-Asndm, 7.
1 Kitdb al-Asndm, 8, 58.

3 Reste, esp. 10-64; Arabie, esp. 101-79; see below, IV pp. 138-40.
* Wellhausen, Reste, 2j8; see below, III pp. 93-9, IV pp. 122-7.

5 Lammens, Arabie, 152; Wellhausen, Reste, 145.

6 Kitdb al-Asndm t 7.

' Wellhausen, Reste, m. 8 Kitdb al-Asndm, 26, 39, respectively.

i8

QURANIC STUDIES

feature of scriptural syntax is the predominantly liturgical use of the
emphatic particle iyya, with pronominal suffix designating the object of
worship, fear, service, etc., before the verbs: 'abada (Q. i : 5, 34: 40, 10:
28, 28: 63, 2: 172, 16: 114, 41: 37, 29: 56), da'a (6: 41), rahiba (2: 40,
16: 51), ittaqd (2: 41), and ista'dna (1 : 5). 1

The primary function of Quranic supplicatory formulae is the invocatio,
expressed in the tahlil «U!l ^1 aJI V (Q. 37: 35 and variations passim), in the
tasmiya «U)I jT)> (e.g. 5 : 4), in the takbir jS'] <U)I ^"jJj (e.g. 29: 45),
or in the introductory imperative to Surat Ikhlds <U)I ja Ji. As song
of praise (hymn) the invocatio seems likely to belong to a cultic tradition,
related to sanctuary and public worship. 2 Some support for that conjecture
might be found in the relation of the tahlil to sacrificial prescription, i.e.
uhilla in Q. 2: 173, 5 : 3, 6: 145, 16: 115, and glossed in 6: 121 ^~1>
<uL& «U)I The alternative is to postulate an origin in private devotion.

While evidence for a cultic background is more abundant, it has been
transmitted in a manner so overlaid with retroflective exegesis (Nachdich-
tung) as to be nearly unintelligible. Solution to this kind of problem is
often frustrated by the impossibility of isolating a satisfactory prehistory
for what in the exegetical tradition have become highly charged technical
terms. A perfect illustration is provided in Kister's study of tahannuth,
in which the author argues persuasively in favour of the equivalence
tahannuth: ta'abbud, accepting the latter as designation of cultic practices
connected with the pre-Islamic sanctuary at Mecca. 3 On the other hand,
the notion of informal, private devotion contained in the Hebrew cognate
(?) tehinnah is rather older than Kister, following Goitein, appears to
accept, though that in itself does not of course justify identification of the
two terms. 4 It is merely that references adduced to support the equivalence
tahannuth: ta'abbud jtabarrur are drawn from a literary tradition which
stresses the exclusively Arabian sources of Islam. 5

The fragmentary character of Muslim scripture can nowhere be more
clearly observed than in those passages traditionally described as narrative.
These consist in fact not so much of narrative as of exempla, of the sort
alluded to in the Qur'an itself as 'signs' (dydt), and hardly qualify even for
the epithet 'legend'. 6 Exhibiting a limited number of themes, the exempla
achieve a kind of stylistic uniformity by resort to a scarcely varied stock of
rhetorical convention. An instance is the serial employment of the pre-
sentative wa-idh (cf. Hebrew nnsi) in Surat Baqara, e.g. verses 30, 34,

1 Cf. Fischer, 'Zunburija', 153-5.
1 See Koch, Growth, 161-6: tehillah.

3 'Tahannuth', 323-36.

4 Kister, op. cit. 231 n. 51 ; see Elbogen, Gottesdienst, 73-81, and 265 : tahanun.

5 See below, IV pp. 122-7.

6 Cf. Horovitz, Untersuckungen, 7 ff.

REVELATION AND CANON

19

49-S 1 . 53-5. 5 8 > to- 1 * 6 3> 6 7. 72, 83-4, 93, 124-7, relieved by idh alone
in 131 and 133, and not noticeably different from constructions with
wa-idhd in the same contexts, e.g. 76, 91, 117.* While the effect is tedious
in the extreme, by means of this deictic formula, of far higher frequency
than any other, a number of disparate topics, abruptly introduced and as
abruptly dismissed, is mechanically linked. Dismissal is commonly effected
by the locution SjV dllS <J <jl ( e -g- Q- 2: 248), with minor variations,
e.g. 'ibra{$: 13), tadkkira(6g: 12). 2

These basic 'narrative' conventions are supplemented by others, which
may be distinguished as interrogative, imperative, and simple declarative,
exhibiting thus the grammatical types already encountered for apodictic
and supplicatory formulae. Since, indeed, the so-called narrative sections
of the Qur'an are of essentially symbolic character adduced to illustrate the
eschatological value of the theodicy, it is not surprising to find such
remarkable conformity. Formally interrogative locutions function almost
exclusively as rhetorical questions, e.g. 6 UJ/oj.U liU'f Jjt> preceding
mention of Abraham, Moses, Pharaoh, Thamud, and other representatives
of the umam khdliya (Q. 20: 9, 51: 24, 79: 15, 85: 17, 38: 21, and the
variation tU dii'L jjl 9: 70, 14: 29). A parallel concluding formula may
be seen in 6j^k ^i-^ (Q - 7 : l8 5» 77 : 5°)> wnere hadith must,
as elsewhere, be understood not as Mstoria but as exemplum, as in 34: 19
p&ULi».j. Similarly, naba' and qissa in phrases of formulaic
character signify exempla, e.g. Q. 18: 13 p*c.Lj <iLU J*m> {f* and 12: 3
ua^aiJI j tt1 - 1 tiLU Jlsii ,ycj. A finite verb of the second derived form
nabbda, imperative in Q. 15: 49, 51 and imperfect in 5: 105, 6: 60, 108,
etc., supports that interpretation, as does the concluding formula dUi/dDJ
e.LJI ^ (e.g. Q. 11: 49, 12: 102), where anba' may be seen as equivalent
to ay at (signs). The rhetorical question is also posed in the locutions
O'/JI J (e-g- Q- 4: 44> 4°: 6 9> 14: J 9> ™' l8 ). 61/JI by. (e.g.
*3 : 4 1 , 34 : 9). and u^l J ,«Jjf ( e -g- 3°: 9. 35-' 44)- Imperative

formations with the verbs dhakara and dhakkara (recall, remind) serve
also to introduce exempla, as in the series <_>k50l J (Q- 19: 16, 41,

51, 54, 56) and Ujuc J*"}\j (Q. 38: 17, 41, 45, cf. verse 48), but especially
14: s iJi\ j»L>L pt> Similarly, the imperative construction fanzur kayf

(see how, consider what) with 'dqiba (destiny, Q. 7: 103), with ayat (signs,
6 : 46), or with amthal (similes/parables, 17 : 48), may introduce or conclude
mention of the exempla which characterize this very allusive version of

1 See N61deke 4 Zur Grammatik, 108-9; Reckendorf, Verhaltnisse, 475-80.

2 Horovitz, op. cit. 6, 13-4.

20

QURANIC STUDIES

Heilsgeschichte. Their paideutic function is explicit in such formulae as
j*$J (strike for them a parable, Q. 18: 32, 45, 36: 13), Jiij

e.Li ^Jtfr (recite to them the example, 5 : 27, 10: 71, 26: 69), and obi
U^kJ aJU 1 (these are the signs of God which we recite, 2: 252, 3: 108,
45 : 6). Functionally related to the imperative constructions is the use in
Surat Saffat of the phrase saldm 'aid (benedictio) to introduce the figures
of Noah, Abraham, Moses, Aaron, etc. (Q. 37: 79, 109, 120, 130).

Despite the uniformity of introductory and concluding formulae, it is
hardly possible to regard these as fixing the limits of recognizable narrative
units. 1 It is even more difficult to reconstruct from the elliptically drawn
exempla themselves a form which might plausibly represent an earlier
stage of transmission. Quranic allusion presupposes familiarity with the
narrative material of Judaeo-Christian scripture, which was not so much
reformulated as merely referred to. Narrative structure on the other hand,
absent in the text of scripture, emerged in the literature of haggadic
exegesis, in which the many lacunae were more or less satisfactorily
filled. 2 Determination of a likely Sitz im Leben for the Quranic forms
turns, then, upon the possibility of regarding the exempla as originally
distinct pericopes of essentially homiletic purpose. There are, of course,
several situations at least in which such pericopes might or could have
originated. But taken together, the quantity of reference, the mechanically
repetitious employment of rhetorical convention, and the stridently
polemical style, all suggest a strongly sectarian atmosphere, in which a
corpus of familiar scripture was being pressed into the service of as yet
unfamiliar doctrine. The implications of that hypothesis for a textual
history of Muslim scripture remain to be examined.

The Quranic masorah is traditionally restricted to a series of more or less
standard deviations from the canonical text. 3 Such material as was alleged
by some to have a claim to canonical status, but which in the event was not
included in the definitive text, tended to be relegated to discussion of
specifically halakhic questions or to argument arising out of sectarian
interests. 4 The problem of 'variants' can be usefully approached by
distinguishing between variant readings, the proper concern of the
masoretes, and variant traditions. In the Muslim exegetical literature the
latter were explained, or evaded, by reference to the chronology of reve-
lation, by means of which unmistakable repetition in the Quranic text

' Cf. Koch, Growth, 1 15-18; Torrey's sociological observations are of little value for
a literary analysis, Foundation, 105-16.
1 See below, IV pp. 127-9.

3 See Jeffery, Materials; and GdQ iii. ♦ GdQ i, 234-61, ii, 33-8, 81-112.

REVELATION AND CANON

zi

could be justified. Versions of the chronology, together with traditions
relating to the modes of revelation, have been considered adequate
criteria for describing the collection and preservation of that text by the
Muslim community. 1 But variant traditions are present in such quantity as
to deserve some attention in a description of the process by which revela-
tion became canon. Unlike the minutiae to be gained from variae lectiones, z
analysis of variant traditions will not support the theory of an Urtext nor
even that of a composite edition produced by deliberations in committee,
both of which may, not surprisingly, be traced to Rabbinic Vorlagen. 3
Such analysis indicates, rather, the existence of independent, possibly
regional, traditions incorporated more or less intact into the canonical
compilation, itself the product of expansion and strife within the Muslim
community. Now, that kind of variant is found for most of the exempla
adduced to illustrate the umam kkaliya theme, of which those relating to
the mission of the prophet Shu'ayb offer a typical instance. It may be
conceded that the kind of analysis undertaken will in no small measure
determine the results. For the Quranic material pertaining to Abraham,
the studies of Beck and Moubarac were developed from a wholly arbitrary
adherence to the traditional chronology of revelation and ended in a
'historical' survey of the prophet's 'changing attitude towards the patri-
archs'. 4 Demonstration of the 'historical development of Abraham in the
Qur'an', for Moubarac the evolution of a composite figure out of an
originally dual image, required not only a verifiable chronology of reve-
lation but also the structural unity of the canon. Both were asserted;
neither was proved. 5

A literary analysis, as contrasted with a historical or theological one, may
properly disregard such criteria. For Shu'ayb the scriptural account exists
in three complete versions: Q. 7: 85-93, 11 : 84-95, 2 & : 176-90; and in
abridged form (introductory and concluding formulae only) in 29: 36-7.
Between the three versions an internal relationship may be set out as
follows.

A

I And to Midian their brother Shu'ayb. (7 : 85)

II (He said) if some of you believe in that with which I have been sent, and
others of you do not, then have patience until God judges between us, for
He is the best of judges. (7. 87)

1 See below, pp. 33-46. 2 See below, IV pp. 203-8.

3 Cf. Eissfeldt, Einleitung, 695-707; and see below, pp. 45-52.

4 Beck, 'Abraham', 73-94; Moubarac, Abraham, esp. 31-50, 91-5, 163-75.

5 Moubarac, Abraham, 139, admitted the difficulty of eliciting from the document of
revelation a 'theology of history' as contrasted with 'reports of a historical nature':
'Ou encore, il y a dans le Coran une histoire de la religion plutdt qu'une religion de
rhistoire'.

22

QURANIC STUDIES

III My people, worship God, for you have no other than Him. And a sign has
come to you from your Lord. Give full measure and weight, and do not
cheat your fellows of their property, nor work evil in the land, now that
matters have been put right. There will be benefit to you if you have
faith. So occupy not any path, seeking to make it devious, and threat-
ening and barring from the way of God him who believes. (7: 85-6)

IV But remember that you were few and He increased you, and consider the
destiny of those who work evil. (7: 86)

V And the assembly of those of his people who regarded themselves as
authorities said: We will expel you from our midst, Shu'ayb, and those
who believe with you, unless you return to our law. (7 : 88)

VI He replied : And if we do not wish to ? We should be guilty of denying God
were we to return to your law after He delivered us. It is not for us to
return unless our Lord God wishes it. His wisdom comprehends all
things, upon God do we rely. Lord, judge in truth between our people
and ourselves, for You are the best of judges. (7 : 88-9)

VII And the assembly of those of his people who did not accept said : If you
follow Shu'ayb then you are lost. (7 : 90)

VIII So disaster overtook them and they were left prostrate in their place.
Those who rejected Shu'ayb were as though they had not been, those who
rejected Shu'ayb were indeed lost. (7: 91-2)

IX And he turned away from them saying: My people, I brought you
messages from my Lord and advised you, so how shall I have pity on a
people who does not accept? (7: 93)

B

I And to Midian their brother Shu'ayb. (11 : 84)

II (He said) My people, do you not see that I bear a sign from my Lord,
Who has provided well for me. I do not wish to oppose you in that which
I forbid, but only to put things right so far as I can. And my success lies
with God upon Whom I rely and towards Whom I turn. (11 : 88)

III My people, worship God, for you have no other than Him. Give not
short measure and weight, for I see you in wealth and fear for you the
punishment of the last day. My people, give full measure and weight in
equity, and do not cheat your fellows of their property, nor work evil in
the land. That which abides with God will be a benefit to you, if you have
faith. For I cannot preserve you. (11 :84~6)

IV My people, let not rejection of me make you guilty, lest that afflict you
which afflicted the people of Noah or of Hud or of Salih, nor are the people
of Lot irrelevant. (11 : 89)

REVELATION AND CANON

23

V They replied: Shu'ayb, does your prayer command that we abandon that
which our fathers worshipped or that we (do not) conduct our affairs as
we wish? But are you not clement and righteous? (11 : 87)

VI (He said) Ask forgiveness of your Lord and turn towards Him. My Lord
is merciful (and) loving. He said: My people, is my family dearer to you
than God, for you have cast Him behind you. My Lord understands that
which you do. (11: 90, 92)

VII They replied: Shu'ayb, we do not understand much of what you say
and regard you as without support. Were it not for your family, we would
stone you, for you have no power over us. (11 : 91)

VIII And when Our decree was uttered We delivered Shu'ayb and those who
had faith with him, as a mercy from Us. And disaster overtook those who
had done wrong and they were left prostrate in their place. As though
they had not been, Midian perished as Thamud had perished. (11 : 94-5)

IX (He said) My people, do in your position (as) I do. You will know him
to whom a grievous punishment comes, and who it is who lies. And
watch, for I watch with you. (11: 93)

C

I ...

II (Shu'ayb said) I am a messenger entrusted to you. (26: 178)

III Thus said Shu'ayb to them : Will you not fear ? Fear God and obey (me).
For that I ask of you no reward. I have no reward but with the Lord of
the universe. Give full measure, and be not of those who cause loss.
Weigh in straightforward equity. And do not cheat your fellows of their
property, nor work evil in the land. (26: 177, 179-83)

IV And fear Him Who created you and those who went before. (26: 184)

V They replied: You are one of the possessed. And are only a mortal like
ourselves, and we think you a liar. But if you speak the truth, then let
some part of heaven fall upon us. (26: 185-7)

VI ...

VII ...

VIII The men of the copse rejected the messengers. Thus they rejected him,
and the punishment of the last day, a momentous day, overtook them. In
that there is a sign, but most of them have no faith. (26:176, 189-80).

IX He said : My Lord knows best what you have done. (26: 188)

The isolation and order of components proposed here, though not
always coincident with verse juncture and sequence, correspond to a
scheme widely attested in the literature of prophetical expression. 1 From
the analysis version A emerges as the most coherent and version C the

1 See Westermann, Grundformen, 79-82; Koch, Growth, 190-4.

24

QURANIC STUDIES

least. Absence in the latter of element I, the formula of commission
(Beauftragung), and the presence there of the simple presentative idh qala
(thus he said: at the beginning of element III) suggest a primitive and
probably earlier formulation. The commission itself, which appears not
only in versions A and B, but also in the summary account at Q. 29: 36-7,
is a stereotype phrase employed exclusively for the non-Biblical 'prophets' :

-Oil IjJufcl fjS L <JU . . . . tJb (Q- 7: 6 5> "^S — Hud to

'Ad; 7: 73, 11 : 61 — Salih to Thamud; 7: 85, 11: 84, 29: 36 — Shu'ayb to
Midian), and might well be thought an editorial interpolation designed
precisely to introduce reports of prophetical missions, themselves cast in
informal dialogue.

Element II, the corroboratio (Botenformel), is not necessarily out of
place in the canonical sequence of A and B, but is probably nearer its
original position in C, where it introduces the substance of the message,
which is element III: the diatribe or accusation {Scheltwort/Anklage). 1
Only in version B is the diatribe not immediately followed by element IV:
the threat or prediction of disaster (Drohwort/UnheilsankuTidigung). 2 Also in
version B that prediction is specified as the destiny of those who rejected
the warnings of Noah and Lot, and of the Arabian prophets Hud and
Salih, while in versions A and C allusion to the fate of the umam khaliya is
general. Thus version B, rather than A, might represent the final shape of
the tradition, a possibility strengthened by the form if not the position of
element II : in C merely rasul amin, in A billadM ursiltu bihi, and in B
bayyina min rabbi (which however occurs, slightly varied, in element HI of
version A: bayyina min rabbikum).

In elements V, VI, and VII, which constitute the altercation, dialogue is
prominent. Version C contains only element V, the contestation (Bestrei-
iung), elaborated in A and B to include a counter-argument or justification
(Begriindung), and a third component (element VII) which might be
described as the resignation and which concludes the dispute. Here,
however, version A exhibits a context more even and consistent than B, as
well as a more sophisticated argument. Element VIII signals reversion
from dialogue to narrative, in order to describe rejection of God's message
and fulfilment of the threat (element IV), the dominant motif of the
Quranic exempla. The position of element IX, the epilogue or final assess-
ment, is logical only in version A. From the canonical verse sequence of
versions B and C, and from its absence in Q. 29: 36-7, one might conclude
that it was a late and optional embellishment.

As set out in the diatribe (element III), the primary transgressions of
Shu'ayb 's people were three : failure to observe fair practice in what appear
to be basic mercantile transactions jl>J I J Hjijl (cf. especially Q. 83 :

1 Westermann, Grundformen, 46-9.

1 Westermann, loc. cit.

REVELATION AND CANON a S

2-3, but also 6: 152, 17: 35, and 12: 59, 88; zndpassim in the negotiations
between Joseph and his brothers) ; lack of equity in questions of property
^e-Uil (j-UI lj^?eJ Vj (cf. Q. 2: 282, 11: 15); and apparently a general
inclination to damage and injury ^jVI J IjJu^ii' V or < J>/i\ J IjiUi" V
^ X*Ju (the latter, attested otherwise only at Q. 2 : 60, evidently restricted
to the Shu'ayb traditions). From indications such as these it is of course quite
impossible to identify either the people or the prophet. 1 Other allusions are
similarly unhelpful, e.g. the concept of 'reform' (isldh) in Q. 7: 85 and
1 1 : 88, but not in version C, or inclusion of the men of the copse among the
ahzab in Q. 38: 13 (cf. also 15: 78) or with others of the umam khdliya in
50 : 14 (cf. also the Midianites in 9 : 70). But Quranic references to Midian
belong as often to the Mosaic traditions, and the story of Shu'ayb must be
regarded as a parable, its exemplary function of greater moment than its
improbable historicity. Admonitions on equity in the matter of weights
and measures represent a Pentateuchal motif, developed in the prophetical
literature (Leviticus 19: 35-6, Deuteronomy 25 : 13-16, andcf. Amos 8: 5,
Micah 6: 11). The Shu'ayb traditions exhibit little by way of historical
development but ample evidence of literary elaboration, drawn from
recognizable and well-established types of prophetical report. Such
elaboration is characteristic of Muslim scripture, in which a comparatively
small number of themes is preserved in varying stages of literary achieve-
ment.

The effect of variant traditions may be differently illustrated, and
assessed. Majority opinion in the exegetical literature accepts that the
passage Q. 55 : 46—77 describes four gardens, whose identification as stages
of celestial reward for the faithful became a significant exercise in eschato-
logical writing. Early on, but with little effect upon the verses as source of
doctrinal and allegorical speculation, a dissenting view was recorded,
namely, that the dual form jannatdni in Q. 55: 46 was demanded by the
scheme obtaining there for verse juncture, but in fact represented the
singular janna, as in Q. 79: 41 ax*, ilj! J Jls; o\^>. Ajj J.IL. iJU. ^Jj
iL^UJI J*^ I jis (£jt<J1 ^ aa*JI jU «0j5f~. z Applied also and for the
same reason to Q. 55 : 62 that stricture would halve the total number of
gardens, a result confirmed by examination of the parallel descriptions of
the two pairs. Each contains sixteen verses (i.e. Q. 55: 46-61 and 62-77),

half of which consist of the recurring phrase <jk.i5Cj U5oj e-il
employed thirty-one times in the entire sura but nowhere else in the text

1 Cf. Geiger, Was hat Mohammed, 170-7; GdQ i, 151 n. 9; Horovitz, Untersuchungen,
93-4, 119-20, 138; Torrey, Foundation, 71; Speyer, Erzahlungen, 251-4.

1 Farra', apud Suyuft Itqdn iii, 299; employing the same criteria but ignoring the
masoretic tradition, ,Muller, Reimprosa, 132-3, reached a similar conclusion; cf. BSOAS
xxxiii (1970) 389-91.

26

QURANIC STUDIES

of scripture (cf. the related locutions in Q. 7: 69, 74, 53: 55). Formally, its
function is to stress the dual inflexion characteristic of this passage.
Structurally, it produces the effect of a litany, and its similarity to oVlS? 1 ? "O
VTOn which performs the same function in Psalm 136 has been noted. 1
I should like here to insist upon the term litany rather than refrain. The
role of the latter in the Qur'an and elsewhere, is that of concluding
formula, which does not adequately describe employment of the device in
this passage. 2 There the construction is based upon eight uniformly short
verses, which are those of the second description (Q. 55 : 62-77), expanded
formally and conceptually in the somewhat less uniform verses of the first
description (Q. 55: 46-61), the whole enveloped and punctuated by the
element: Then which of your Lord's bounties do you deny? In the order
of the canon the two sets of eight verses are as follows.

A

46. And to him who fears the presence of his Lord, two gardens.
48. Of many varieties/colours.
50. With two springs flowing.
52. Of every fruit two kinds.

54. Reclining on beds with linings of brocade, the fruit of the two gardens
near by.

56. With chaste maidens, untouched before by men or jinn.
58. As though of ruby and pearl.

60. Can the reward for goodness be other than goodness?

B

62. And besides them, two gardens.

64. Of deep green.

66. With two springs bubbling.

68. With fruit and palms and pomegranates.

70. With pure (and) beautiful ones.

72. Pure white, reserved in chambers.

74. Untouched before by men or jinn.

76. Reclining on green cushions and fine carpets.

The contrapuntal scheme is not perfectly executed, but sufficiently so to
have prompted Zamakhshari to qualify the descriptive components of
version B as just inferior to those of version A. 3 That implied of course

1 Hirschfeld, Researches, 73 n. 20, though of the caique I am less certain than the
author; cf. also Speyer, Erzaklungen, 29.

2 Ibn Qutayba described the formula as a marker (fdsila) between each pair of divine
favours, Ta'wil, 181; cf. GdQ i, 42; Koch, Growth, 96; Eissfeldt, Einleitung, 71-3:
Strophenbau.

5 Kashshdfiv, 454: ad min duni hima in verse 62.

REVELATION AND CANON

27

acceptance of the canonical order of the two descriptions, but from the
same evidence it could be argued that version A represents an elaboration
of version B, both by rhetorical device and exegetical gloss. Whether the
embellishment is to be understood as purely literary or as a reflex of what
may have been the liturgical function of these verses, is difficult to deter-
mine. If a cultic context can be envisaged, it would seem that the descrip-
tions of paradise were recited in inverse order to the canon. More likely,
however, is juxtaposition in the canon of two closely related variant
traditions, contaminated by recitation in identical contexts, or produced
from a single tradition by oral transmission. Imagery based on a pair of
gardens appears elsewhere, e.g. Q. 34: 15-16 and 18: 32-3, though in the
latter instance the parable is weakened by defective syntactical distribution. 1
In neither passage did the imagery provoke eschatological speculation,
whereas Quranic garden imagery based upon the singular janna (e.g. ashdb
al-janna Q. 2: 82 and passim; and the parables 2: 265-6, 68: 17-33)* or
the plural jannat (e.g. the formula jl^j "if I l^^J ^ <£j*Z oLa. Q. 2: 25 and
passim; jannat 'adn 9: 72 and passim; jannat al-na'int 10: 9, and jannat
wa-'uyun 51 : 15), did provide the traditional points of departure for such.
It may well be that the dual jannat ani of Q. 55 : 46 and 62 was, implicitly,
never understood as anything but singular.

From Zamakhshari's pun on min duni hima (Q. 55 : 62) it is legitimate to
infer a consciously formulated correspondence between the descriptions in
verses 46-60 and 62-76. Whether that correspondence is exclusively
rhetorical, or rhetorical and exegetical, will depend upon the more general
problem of recognizing in the canonical text such phenomena as paraphrase,
gloss, and interpolation. That problem was adumbrated by Fischer in his
analysis of Q. 101 : 8— 11, where the exegetical function of the locution ma
adraka ma (how do you know what ... is?) is convincingly demonstrated. 3
Reverberation of the exegetical equivalence hdzoiya: abyss jhell&re found
expression elsewhere, e.g. in an elegy ascribed to Waraqa b. Nawfal
bjU <L)I J I liL*. / >"UJI; 4 and an ironic utterance attributed to 'Abid b.
Abras, describing his narrow escape from eternal damnation fj\ iJ ^-sU
SjjI^JI cii-J, s both emphasizing, incidentally, the triptosy of its scriptural

1 Cf. Speyer, Erzahlungen, 433-4; and below, IV p. 128.

2 Horovitz, Untersuchungen, 11.

3 'Eine Qoran-Interpolation', 33-55 ; though it is easier to identify than to date inter-
polations, the objections to Fischer's proposal articulated by Barth and Torrey are, in
my opinion, unacceptable: references in Paret, Der Koran, 518-19; and O'Shaughnessy,
'Seven Names', 449-51.

4 Ibn Hisham, Sira i, 232 line 5.

5 Ma'arri, Risdlat al-Ghufrdn, 178 top.

28

QURANIC STUDIES

occurrence. 1 That the phrase ma adraka ma is of hermeneutical, not merely
lexical, moment is evident from its thirteen occurrences in the Qur'an where,
in addition to such hapax legomena as hawiya (Q. 101 : 10), saqar (74: 27),
'illiyun (83: 19), etc., expressions like yawm al-din (82: 17-18), yawm al-
fasl(yj: 14), and 'aqaba (90: 12) are glossed. But traditional accommo-
dation of the phrase was designed to expel doubts about just whose utter-
ance it represented: God's, prophet's, or commentator's. Raghib Isfahan!
declared, for example, that whenever God preceded mention of any topic
with ma adraka ma, He explained it, but that when He employed the
expression mayudrika, the matter was left unexplained. 2 The fact that the
two formulae are anyway not functionally comparable is less relevant than
the manner in which ma. adraka ma was there unequivocally assumed to be
the word of God. The ma yudrika construction, always completed by
Walla (how do you know whether?), is predicative rather than explicative
(Q- 33-' 63, 42: 17, 80: 3), as Isfahan! inadvertently observed.

Some instances of apostrophe might be thought to signal commentary to
the text of revelation, e.g. in Q. 16: 51 Ljl/^l IjIstcS V <UJl <Jlij

jj-Ajli t£UU/Jb.lj oVt ja where the middle term in the third person
contrasts with the imperative construction as a whole: 'God said: do not
take unto yourselves two gods/for He is only one god/so have fear of Me
(alone)'. Such, in my opinion, is not to be confused with a passage like
Q. 25 : 45-56, in which mention of God alternates between first and third
person, exhibiting the abrupt transition characteristic of so much of
Quranic style, and indicating not juxtaposition of text and commentary
but rather, of fragments from independent traditions. That the apo-
strophic qui must in those contexts impossibly predicable of God be an
interpolation has been proposed; similarly the locution min allah after
mawthiq in Q. 12: 66, 8o, by means of which that otherwise unattested
designation of covenant could be unambiguously related to mithaq.

In the examples of variant traditions adduced above, those describing
the mission of Shu'ayb might be held to contain traces of exegetical
activity. Element I (formula of commission) appears in versions A and B
only, and may represent a gloss of the introductory idh qala of version C.
Element IV (threat/prediction of disaster), in A and C a general reference
to the umam khaliya, becomes in B specific enumeration of the fates of
those who rejected Noah, Hud, Salih, and Lot. On the other hand, elabor-
ation of element V in version C to the series V— VII (altercation) in versions
A and B is more likely to be evidence of independent origins. Now, in the
passage containing parallel descriptions of the two gardens, the precise
relationship between corresponding elements is rather more problematic,
since, as I have intimated, the possibility of proliferation from a single

1 Fischer, 'Eine Qoran-Interpolation', 43-4.

2 Apud Suyuft Itqan ii, 139.

REVELATION AND CANON

29

tradition cannot be ruled out. 1 However, a tendency towards clarification
as well as conceptual enhancement is discernible in a comparison of verse
64 (mudhammatani) with verse 48 (dhawata afndnin: interpreting afndn as
alzvdn), 2 of verse 66 (nadddkhatdni) with verse 50 (tajriydni), or of verse 68
(fdkikatun wa-nakhlun wa-rummdnun) with verse 52 (min kulli fdkihatin
zawjdni). Thus set out the evidence would appear to support the hypo-
thesis of an earlier composition for version B, though the remaining parts of
the two descriptions suggest other possibilities. For example, subsumption
in verse 56 of all the elements of verses 70, 72, 74 might indicate originally
distinct traditions, as could the presence in version A of verses 58 and 60,
both without obvious counterparts in version B. But I am inclined to see
in verse 58 (al-ydqut wal-marjdn) an elaboration of hur in verse 72, and to
interpret verse 54 as a conscious reformulation of the clumsier imagery of
verse 76. A paraphrase of version B in version A might none the less be of
liturgical, rather than exclusively exegetical origin. The rhetorical question
of verse 60 exhibits a common Quranic device, already noticed, and serves
to stress the eschatological content of the entire passage. As such it would
belong to the exegetical tradition, the concern of which it was to underline
the relation of this particular instance of double garden imagery with the
concept of celestial reward elsewhere illustrated by the forms janna and
ianndt.

I have drawn attention to the allusive treatment in the Qur'an of the
schemata of revelation. The style is best observed in the so-called 'narra-
tive' passages, of which the Shu'ayb traditions offer typical illustration.
Indeed, each of the themes representative of the Quranic theodicy ex-
hibits a recognizable literary type in a state of suspended development.
A degree of uniformity was achieved by recourse to a limited number of
rhetorical conventions. Specifically stylistic problems provoke other more
general ones relating to assimilation and the mimetic process. For example,
traces in the retribution imagery of election and remnant traditions associ-
ated with the salvation history of Israel distinguish the role of that nation
in the theodicy from those of the umam khdliya, themselves elected but
vanished without survivors. A consequence of that distinction is the purely
exemplary function of parables containing the accounts of non-Biblical
prophets, and a concomitant failure to assimilate Arabian elements to the
Judaeo-Christian legacy. An example of partial assimilation may be seen
in the imagery developed round the concept 'days/battles of God'. Not
unexpectedly, the injunction wa-dhakkirhum bi-ayydm allah in Q. 14: 5
is addressed to Moses. But elsewhere the terms ayyam and ayydm allah
are of general application, and the possibility of contamination by the

1 Cf. Koch, Growth, 90-1.

2 Zamakhsharl, Kashshdfiv, 452 ad loc.

3°

QURANIC STUDIES

autochthonic ayyam al-arab has been noted. Other instances of more
and less successful assimilation to the native linguistic tradition are provided
by the Quranic imagery relating to angelology and resurrection.

Anonymous and collective reference in Muslim scripture to angels
(malakjmala.'ika) is abundant, relieved very rarely by specific designation,
e.g. Q. 2: 98, 102. 1 In the exegetical literature the vague and general
became explicit and specific (ta'yiri), and a series of dogmas elaborated
with reference to the role and rank of angels in the hierarchy of creation. 2
The function of angels as mediators between God and men, and by ex-
tension emissaries and agents of revelation, is a significant component of
Islamic prophetology (cf. e.g. Q. 17: 95, 22: 75, 35: 1, 41: 14). 3 Terms
other than malak were in the exegetical tradition subsumed under this
rubric, as for example, zabdniyya in Q. 96: 18, interpreted by Zamakh-
shari as 'angels of retribution' {malaikat al-adhab). 4 Rather less arbitrary
was application of that procedure to the epithet muqarrabun. In four
passages (Q. 56: 11, 88, 83 : 21, 28), the notion of propinquity to God as
reward for piety is clearly conveyed, expressed in 56: 7-1 1 as a tripartite
distribution of benefit, of which muqarrabun represent the highest order.
Twice (Q. 7 : 1 14, 26: 42) the same image is employed in a secular setting,
promise of reward being the utterance of Pharaoh. Thus the transfer of
function achieved by derivatives of the root q-r-b attested in prophetical
and Rabbinic literature is reflected in Quranic usage, though I should be
reluctant to interpret the courtly environment of Q. 7: 114 and 26: 42 as
exhibiting a genuine Sitz im Leben for the Arabic metaphor. 5 As in the
imagery associated with the messenger as both royal and divine agent, that
relating to admission into the royal and into the divine presence may be
thought a caique of the semantic evolution already complete in Biblical
Hebrew, e.g. Esther 1: 14 as contrasted with Jeremiah 30: 21.

In two further Quranic passages (3: 45 and 4: 172), both traditional
sources of Muslim Christology, the content of muqarrabun is extended to
include not only an elite among the saved, but also Jesus and certain of the
angels. Express reference in Q. 4: 172 to malaika muqarrabun became in
the exegetical tradition the occasion of another and different link with
Biblical imagery. ZamakhsharT identified these 'angels drawn near' with
the cherubim round the throne of God 'like Gabriel, Michael, Israfil, and
others of their rank' ^ij^S~ ^J>y>i\ <Jj^. ^jJI Oj^jjSOl iSo^LJI ^aj
J Cr>J Jjj'j-'J Js*^J- 6 In Arabic lexicography approximation

1 See Wensinck, Creed, 198-202 and references ; also Vajda, EI, s.v. Harut wa-Marut.

2 Wensinck, loc. cit; Geiger, Was hat Mohammed, 80-2; for ta'yin, see below, IV
pp. 135-6. 3 See below, II pp. 55, 61-3.

* Kashshdf iv, 779 ad loc.

5 Speyer, Erzdhlungen, 266, 299-300; von Rad, Theology ii, 218-9.

* Kashshdf i, 596 ad loc.; cf. Speyer, op. cit., 26—7 for Rabbinic parallels; and Goldziher,
'Polemik', 371, for kanbiyyun and Israfil in Arabic versions of Genesis 3: 24.

REVELATION AND CANON

3i

of the roots q-r-b and k-r-b was seen to conform to a series of semantic con-
clusions drawn from the behaviour of morphemes of intensity (mubdlagha),
by which karubiyyun must represent those of the muqarrabun 'nearest the
throne of God' and, hence, the principal of the angels, among whom were
Gabriel, Michael, and Israfil. 1 The linguistic laws by which that accom-
modation of a non-linguistic caique was justified are only partially
attested: the measure fa'ul and the affix -iyy may be so interpreted, but
not the relation kjq. z Juxtaposition of three originally unrelated elements :
the pious, the archangels, and the cherubim, is of course logically un-
satisfactory, but for the assimilation of Biblical exegesis to the Arabic
linguistic tradition, an instructive example indeed.

The lexical range of Quranic resurrection imagery is rather wider. The
primary concept is that of resurrection as a second creation, derived from
the vocabulary appropriate to the acts of God in nature. 3 Typical formu-
lations <Jjf pTbsU LT joJ (Q. 18: 48), pTusJii i^u

iSj£-\ ojlj p^a-jSej L^Uj Xj6 (20: 55), 'i Jjt Ut£j| IfejSeJ Ji

(36: 79), and oXjo p$ (jliJI jJuj «U)I Js (10: 34 and passim). Express

reference to hhalq jadid (fresh creation) occurs consistently in a context of
altercation about the promised resurrection, e.g. Q. 13: 5, 17: 49, 98,
32: 10, 34: 7 (variation hhalq akhar in 23: 14), and in 50: 15 paired with
hhalq awwal (first creation). With derivatives of such verbs as ansha'a (in
Q. 36: 79, above), anshara, and akhraja (signifying erect, elicit, evoke),
imagery designed to convey production of life out of death is varied and
extended: nasKa ula (56: 62), nasKa ukhra (53: 47), nash'a akhira(z<):
20), insha' ($(>•. 35), nushur (25: 3, 35: 9), khuruj{^o: 11, 42), etc. Of highest
frequency, however, are constructions based on the antithesis hayy : mayyit,
occurring with akhraja in Q. 6: 95 and 10: 31, but mostly with finite verb-
forms of the same roots, e.g. Q. 2: 259, 22: 6, 30: 50, and explicitly linked
with hhalq in 2: 164 and 46: 33."* The categorical statement in the song of
Moses riTiSl rp»8 "UK (Deuteronomy 32 : 39) appears verbatim in Q. 2 : 259
■-^•t> ti 5 *' ^'' 5 ^ e n y miuc context of the Biblical passages, reflected in
the liturgical application of DTfln rvnn, may be seen also in Quranic
usage, where the epithet 'bringer of life and death/death and life' is often
introduced by the invocatio, e.g. allah, kuzoa, zva-huwa 'lladhi, etc. (Q.
30: 40, 10: 56, 23 : 80). 6 Related to that particular image is the concept, but

1 Zabidl, Taj al-'Arus i, 454 s.v. karubiyyun.

2 Cf. Vollers, Volkssprache, 11; GVGi, 121-2.

3 Cf. Lehmann-Pedersen, 'Auferstehung', 54-61 ; Wensinck, 'Propheten', 182-4.

4 See the passages adduced in Muller, Reimprosa, 128-9; O'Shaughnessy, Death,
32, 45-6, 50; cf. BSOAS xxxiii (1970) 613-5.

5 Cf. also the third person construction in Q. 53 : 44 with the song of Hannah, 1 Samuel
2: 6, adduced Speyer, Erzahlungen, 445.

Cf. Elbogen, Got'tesdienst, 29, 32, 44; Geiger, Was hat Mohammed, 70, 76-8, 202.

QURANIC STUDIES

not the cognate, of 'dry bones' reclothed with flesh at the command of God
imy m&XS (Ezekiel 37: 1-14), e.g. Q. 23: 14 MySi Like ii^JI LiUei
L*J J.U4J.JI and 36: 78 ^lk*JI ^ Rhetorical emphasis

upon the corporeal aspect of resurrection is exhibited also in the collocation
'dust and bones' (turdb wa-'izam, e.g. Q. 23 : 82), an expression of incre-
dulity rather than a divine reminder of mortality. 1 Quranic expression of
the latter is contained in Q. 22: 5 t_>ly Cr* ^* ^ n Muslim

scripture the fact of resurrection is presupposed : it is the point of departure
for polemic rather than the result of thematic development, and formula-
tions of doubt are stereotype.

Of somewhat different character are four other terms appropriate to the
Quranic resurrection imagery: ba'th, hashr, raj', and qiydma. The basic
notion of dispatch common to the first three of these is exhibited in
scriptural usage, but is in each case dominated by association with the
symbolism of the day of judgement. Such is achieved by juxtaposition of
finite forms with the substantive yawm (day), e.g. jUJI | »^> c -j (Q. 56: 6,

18), jji«J fjj J! (7: 14. l 5- 3 6 > etc )- je£«JI f-*5 ( J 9 : 8 S)> fjli
pj*. jZ^ci (10: 28, etc.), <UI dy^-ji fjiJ (24: 64, cf. 2: 281). The day
is occasionally specified as yawm al-qiydma, as in Q. 23: 16, 17: 97, but
the ambiguity inherent in ba'atha (as a synonym of arsala: send) may have
effected what appears to be a gloss (hayyan: alive) in Q. 19: 15 and 33
C* cUjoI/uU*^ fjjj. Similarly Q. 22: 7 jj-SJI <j ^ <±-*-j «Ull jlj where
the locution 'those in their graves' restricts the semantic range of yab'ath
(sends, hence resurrects) in a manner appropriate to the exegetical gloss.
The metaphorical value of the image is thus diminished, or eliminated, as
in Zamakhshari's interpretation of the isolated instance of yawm al-khuruj
(Q. 50 : 42 : day of emergence) as 'they emerge from their graves', which,
though rendering yawm al-khurUj unequivocally as 'day of resurrection',
destroys the more sophisticated metaphorical function of akkraja and its
derivatives. 2 The same ambiguity in hashara (gather) and raja'a (return)
is reduced by employment of parallel constructions with ilayhi (to Him),
e.g. ilayhi tuhsharun (Q. 5 : 96, 58: 9, etc.) and ilayhi turja'un (2: 28, 10: 56,
etc.). Status constuctus with yawm and the verbal noun is explicit only for
ba'th £..«J1 \'x& (Q. 30: 56), but the contexts of Q. 50: 44 t ji£J ^

* Ml

j-^j UJui ^£a* dJUi U-l ju, |t g:c si and 86: 8-9 ^ <u».j ^e. aJI
y\j*J\ jJ^J perform a similar service for hashr and raj', respectively.

Of the four terms qiydma is unique, and for two reasons: the verbal
noun itself occurs only in status constructus with yawm, and, with two

1 O'Shaughnessy, Death, 46-9; cf. Speyer, op. cit. 72-3 ad Genesis 3:19.
* Kaskshqf iv, 393 ad loc

REVELATION AND CANON

33

exceptions, finite forms of the root q-zo-m do not explicitly describe the
fact of resurrection. The locution yawm taqum al-sd'a (Q. 30: 12, 14, 55,
40: 46, 45: 27) reflects al-sd'a qd'ima of 18: 36 and 41: 50, in which the
agent is the (final) Hour. Similarly, combinations of the imperfect with
hisab (reckoning) in Q. 14: 41, ashhad (witnesses) in 40: 51, ruh (spirit) and
mala'tka (angels) in 73 : 38. Only in Q. 39: 68 OjJ^i fLi «a liU and 83 : 6
l ^ s *JU!l CjJ (^"Ul |»jju is the act of rising/standing predicated of men.
On the other hand, ratios of frequency and distribution for yawm al-qiyama
are high (seventy times in suras ranging from the second to the seventy-fifth),
and corresponding collocation with derivatives of ba'th and hashr has been
noted. Against the possibility of generation from the finite constructions
yawma yaqumujtaqumu may be set the likelihood of yawm al-qiyama as an
incompletely assimilated borrowing. A caique of Syriac qeyamta, indicat-
ing presumably a loan-translation of Greek anastasis, would produce a
significant polarity within the Quranic resurrection imagery. 1 The major
implication need not, however, be a dual source for the theological concept,
namely DVWn rvnn and avdaraais, 2 but rather, or also, further evidence
of the polemical environment in which the scripture of Islam came into
being.

2. ITS COMPOSITION

Procedures of transmission and preservation demand that the word of God
conform to recognizable patterns of human utterance. From the foregoing
analysis of rhetorical schemata and of variant tradition, exegetical gloss,
and conceptual assimilation, it may be supposed that the Quranic revelation
is no exception to the general rule. But the mimetic process is a complex
one. Isolation of such monotheist imagery as is characteristic of themes
like divine retribution and sign, covenant and exile, indicates the per-
petuation in Muslim scripture of established literary types. And yet, the
merely allusive style of that document would appear to preclude positing
the relationship of figural interpretation (typology) admitted to exist
between the Old and New Testaments. The pattern of fulfilment {figurant
implere) cannot, or at least hardly, be elicited from a comparison of Muslim
with Hebrew scripture. That this is not merely a negative inference from the
absence of an explicit connection of the sort established between the Chris-
tian and Hebrew scriptures ought to be clear from examination of the
Quranic forms themselves, which reflect, but do not develop, most of the
themes traditionally associated with literature of prophetical expression.
If the claim to place the Qur'an within that clearly defined literary tradition
is conceded, it would none the less be inaccurate to describe that document
as exhibiting essentially a caique of earlier fixed forms. The relationship

1 Ahrens, 'Christliches im Qoran', 32. 2 Cf. Black, Aramaic Approach, 281.

4S39C75 C

34 QURANIC STUDIES

is rather more complicated, due at least in part to the origins of Muslim
scripture in polemic. Illustration of that is provided by analysis of the con-
fused and conflicting theories about the manner of its composition.

Muslim views on the mechanics of revelation consist for the most part
of exegetical speculation on the content of Q. 42: 51. That verse may be
set out as follows.

It is not granted to any man ^ ^

that God address him except: ^ ^

(A) directly (A)

(B) or from behind a screen *b-S lt* ( B )

(C) or by sending a messenger X>~J J-yi jf (C)
who utters with His permission *>ty ^fji*
that which He wishes

He is indeed exalted and wise. ^ 4iJ

A primary difficulty in this passage is its implication for a fundamental
dilemma of Islamic theology: revelation as the unmediated speech of God,
or revelation as the prophetical (angelic) report of God's speech. My trans-
lation of element (A) diverges from the consensus doctorum of Muslim
tradition, according to which wahy is synonymous with ilham (inspiration),
the verbal noun of Quranic alhama(gi : 8, a hapax legomenon). 1 It seems
clear, however, from element (C), in which the pronominal components of
fa-yilhiya and bi-idhnihi can hardly share a single referent, that the use here
of awha is 'to reveal/present oneself and, in conjunction with kallama, 'to
utter directly (without mediation)'. That interpretation has the additional
advantage of offering the required degree of contrast between the three ( !)
alternative kinds of theophany. Zamakhshari, drawing upon the imagery of
delegated authority (wakil, rasul), permitted element (C) to be so inter-
preted, but alluded in the same passage to what had become a traditional
link between elements (A) and (C), namely, that the concept wahy pre-
supposed dispatch (irsdl) of a messenger. 2 Now, that Quranic awha may
in some contexts be a synonym of arsala is clear from the very next verse
(Q. 42:52) U^it j-Ujj idLlI Uj»jidUj^(cf. 17:86,41: 12). Application of
the equivalence to elements (A) and (C) of Q. 42: 51, producing ultimately
an interpretation of nearly the same currency as wahy : ilham, may, I suspect,
be traced to the elliptical style of earlier exegetes. Muhammad Kalbl, for
example, glossed 'We reveal to you' with 'We send Gabriel to you with it'
in the sense of 'to inform you of it'. 3 Muqatil b. Sulayman improved upon

1 e.g. Zamakhshari, Kashshaf iv, 233 ad loc. Ibn Qutayba, on the other hand, was
doubtful, Ta'wil, 78, 83. See below, II pp. 58-9.

2 Kashshaf, loc. cit. ; the equation wahy: irsdl may be compared with Zamakhshari's
observations on the messenger formula as derived from royal protocol, see above, p. 12.

3 Tafsir, MS Ayasofya 118, 135' ad Q. 12: 102.

REVELATION AND CANON

35

that method by adding to his own gloss 'God revealed to him' the phrase
'Gabriel came to him and informed him of it'. 1 But that wahy may signify
communication, without recourse to an emissary, is confirmed not only by
Q. 42: 51 but also by the use oiawha in 6: 19, 112, 12: 3, 18: 27, etc. 2

The significance of the tripartite description of the word of God in Q. 42:
5 1 lies in its allusion to the uniqueness of the Mosaic revelation, explicitly
adduced three times by Zamakhshari in his commentary to this verse.
According to that exposition divine communication to all prophets other
than Moses was conveyed by an emissary (mode C); Moses and the angels
alone was/are addressed by God, but indirecdy (mode B) ; the Jewish claim
that Moses had been directly and personally spoken to by God (mode A)
was denied. This report, adduced without authentication, is found also in
Muqatil, and is symptomatic of the polemical atmosphere in which Muslim
views were formulated. 3

The exact nature of that polemic emerges from examination of the
imagery employed for mode (B). Quranic hijab (screen) may be of literal
application (Q. 33 : 53) or metaphorical (19: 17, 38: 32, 83: 15); its function
in 7 146 is eschatological, and in 17 : 45 and 41 :5 it is a reflex of Biblical
masveh/kalymma.* In Q. 42: 51, however, the symbolism is Rabbinic, being
the locution pVnri "HinNB descriptive of the distinction between Israelite
and foreign prophets in their reception of the word of God. 5 And within
the circle of Hebrew prophets the Biblical distinction accorded Moses
(Exodus 33: 11, Numbers 12: 8, Deuteronomy 34: 10) and elaborated
in the Rabbinic tradition, 6 is also and not unexpectedly found in Muslim

exegesis, e.g. adQ. 2: 253, 4: 164 (L-1SJ <UJ| f »15j), 7 :I 43~4> 28: 30,
where it reflects a transition from mode (B) to mode (A). 7 Biblical attestation
of the unique relationship between God and Moses found a second expres-
sion in the criterion of angelic mediator, essential to all but the Mosaic
revelation and, as mode (C) of Q. 42 : 51, of considerable significance in the
development of Muslim prophetology. 8 The unmediated theophany of the
Pentateuch and earlier prophets was recast in prophetical literature proper,
to which the Qur'an may be reckoned, by recourse to the messenger formulae
and the divine imperative. 9 Quranic wahy in its final form was an inclusive
concept, expanded by Suyuti to cover not only the express utterance of

1 Tafsir, MS. H. Husnii 17, 172" ad Q. 18: 86.

2 For wahy in the profane tradition, cf. Goldziher, Studien ii, 7 n. 1.

3 Kashshdf iv, 233-4; Tafsir, MS H. Hiisnii 17, 277* ad loc; cf. Geiger, Was hat
Mohammed, 78-80, and Speyer, Erzaklungen, 599-301.

4 See below, II pp. 72-3.

5 Speyer, op. cit. 420; Horovitz, Untersuchungen, 52-3; Jeffery, 'Scripture', 200 n. 34.

6 Cf. Saadya, Kitdb al-Amdndt, 118-25, 132-3; Maimonides, Daldlat ii, chs. 33-4;
Altmann, 'Theory*, 19-21; Speyer, op. cit. 419.

7 Zamakhshari, Kashshdf i, 297 ad Q. 2: 253; Baqillani, I'jdz, 1 5 ; Suyuti, Itqdn i, 129;
Sharif Murtada, Amdli ii, 312 ad Q. 4: 164; GdQ i, 23.

8 See below, II pp. 58-61. » Cf. Koch, Growth, 187; and above, pp. 12-15.

36

QURANIC STUDIES

God but also that which men among themselves perceived to be His
intention (i.e. inspiration). 1 As set out in Q. 42: 51, and shorn of profane
and rhetorical amplification, the notion drew almost certainly upon Rab-
binic formulations of the Mosaic tradition, even to the extent of adopting
in modified form arguments designed originally to demonstrate the pre-
creation of the Torah. 2

Muslim discussion of the mode of Quranic revelation is characterized
by a predominant concern to distinguish it from the manner in which the
Torah was revealed. Expressed in the polarity jumla (integral) : munajjam
(serial), the distinction derived moment from a widely accepted inter-
pretation of Q. 25: 32 i'J^lj aLs- oTjaJI aJ* ijji VjJ jjjJI Jlij
%3 Ji eliljjj italji Aj dU ji". The necessary link between the locution

jumlatan wahidatan (all of one piece) and the Mosaic revelation was pro-
vided by interpreting alladhina kafaru (those who reject/disbelieve) as
reference to the Jews. Ina prophetical hadith traced to 'Abdallah b. 'Abbas
such was the primary interpretation, mushrikun(\>o\ytfazisXs) being adduced
as an alternative. For Zamakhshari the roles were reversed. The spokesmen
are Quraysh or, it is said (zva-qila), the Jews. The contrast may be under-
stood as polemical rather than historical, and exhibits a major theme of
Muslim exegetical literature. 3 Priority of the Jews over Quraysh in that

particular context was early attested, e.g. in Kalbi's gloss to Q. 17: 2 Uj'Tj
iJU-lj aLjs. i'ljjxli LJifrf/(_>b30 1 ^>y>- Quranic proof-texts for this

interpretation, containing reference to the Mosaic tablets (alwdh, e.g. Q. 7 :
144-5, 1 S°> I S4> I 7 I )> were easily found and frequently adduced, as in
Suyiiti. 4 Preoccupation with the precise difference between the two revela-
tions generated a number of near-synonyms for the adverbial munajjaman
(also nujuman): in the second half of Q. 25: 32 the term tartilan, properly
an elocutionary designation {tafiij al-asnan) was often interpreted as a
reference to serial revelation; from Q. 17: 106,^1^)1 ^p. «{ Jti tbs^s UT^sj
"^oyZ »U_yj J* the expressions mufarraqan (separately) and tanzilan
(by descent) could be seen to embody the same concept. 5 Similarly, nujuman,
somewhat arbitrarily related to Quranic (56: 75), became the

point of departure for speculation on the likely extent/capacity of an in-

1 Itqdn i, 128, on the doubtful authority of Zuhri.

* See Ibn Wahb, Burhdn, 139-41 ; GdQ i, 21 n. 2, 46; Horovitz, op. cit. 67-8; and below,
II pp. 83-4.

3 Suyiiti, Itqdn i, 122; Zamakhshari, Kashshdf iii, 278 ad loc; and see below, IV
pp. 122-6.

4 Kalbi, Tafsir, MS Hamidiye 40, i74 v ; Itqdn i, 122-3 ; cf- Geiger, Was hat Mohammed,
151.

s Zamakhshari, Kashshdf iii, 278-9 ad Q. 25: 32; cf. ibid iv, 637 for tartilan in Q. 73:
4; Suyuti, Itqdn i, 117-18.

REVELATION AND CANON

37

stalment (najm), generally agreed to contain about five verses. The notion

of a time-lag may also be elicited from Q. 20: 114 JJ> ^ j^/Jl? J^""

dLJI ^^j* 6? in which the terms quran (publication) and wahy

(communication) are balanced by implicit reference to a period required for
hearing, understanding, and learning the content of each revelation. 1

The antithesis jwnla : munajjatn was further differentiated. Lest the con-
cept of piecemeal revelation be seen to thrown doubt upon the origins of

Muslim scripture, two verses jT^JI *J J^f i£jJI OUa-y (Q. 2: 185)
and j Ju)| 3iLJ J t Uy\ CH (97: 1) were adduced and interpreted to demon-
strate that the Quranic revelation had consisted of two separate processes:
transfer in its entirety to the nearest heaven and thence serially to the

prophet during a period of approximately twenty years tU^JI ^1 Jjj <*j|
^ j^Uc fj L*e1^» dJii Jju J^S e^a-lj iLws- aL) LJjJI. 2 The
chronological expressions in the two verses, Ramadan and Laylat al-
qadr, did not remain exclusive to the Muslim revelation: the scriptures
of Abraham, of Moses, of David, and of Jesus had also been revealed at
ascertainable dates in Ramadan. 3 But it is the Mosaic revelation which
furnished almost alone a point d'appui in the polemical discourse, ex-
hibiting the Rabbinic (and Patristic) view of the origins of the Penta-
teuch. 4 On the other hand, references to Christian scripture in the Qur'an
provoked inevitably a similar description, as in Suyuti, where the inter-
pretation of Q. 3 : 3 stressed that both Torah and Gospel (injil) had been
revealed all of a piece, or in the observation of Ibn Ishaq on the material
contained in the revelation to Jesus ^2>j jl5" L^t ^jAi LJ jl5" Jij
J~*j*VI JaV J~?JVI <j "Ull oaU. Li pij^ ^|. s Despite polemical
reiteration some confusion persisted, ultimately to be clarified or evaded
by resort to rationalizing argument drawn from outside scripture. The
munajjam concept was after all not exclusively Quranic: in his commentary
to the Psalter Saadya explained repetition as the consequence of separate
occasions of revelation to David, much as could be found in the Torah,
itself a product of serial revelation nsru nVifi n"?JO n*ilD. 6

The dual character of the Quranic revelation was seen to be an honour

1 Suyuti, Itqdn i, 124-5; Zamakhshari, Kashshdf Hi, 90 ad Q. 20: 114.

2 Suyuti, Itqdn i, 116; cf. GdQ i, 80; the accusative pronoun in Q. 97: 1 is in Suyutf's
discussion understood to refer to the Qur'an, though Gabriel was occasionally made the
referent, see below, II pp. 61-2.

3 Suyuti, Itqdn i, 120; a variant tradition in the anonymous Kitdb al-Mabdni, 235.

4 Cf. Eissfeldt, Einleitung, 695-8, citing inter alia Talm. Babl. Baba Batra 14b; and
see Goitein, 'Ramadan', 90-110.

5 Itqdn iii, 343 ; apud Ibn Hishara, Sira i, 232.

6 Lauterbach, Saadja, 23 n. 9; Speyer, Erzdhlungen, 423 n. 2, citing Talm. Babl.
Gittim 60a.

3»

QURANIC STUDIES

(tashrif) bestowed by God upon its recipient: by virtue of the integral
transfer of the word to the nearest heaven, Muhammad was made the
equal (taswiya) of Moses, above whom he was subsequently elevated

(tafdil) as a result of the serial transfer ^ U^j I

<dii*J U^l. .die AJIjjj J Ju*J J^fcJIj J. Precise

reasons for the second process were of three kinds: (a) Muhammad, being
illiterate, required time to learn; (b) the Quran, containing abrogating
and abrogated parts, revised formulations, and direct answers to specific
questions, must of necessity have been serially revealed ; (c) conversion of
the Arabian heathen had perforce to proceed by gentle and humane
stages. 1 The extent to which those views reflect doctrinal disputes of a
most significant kind for the historical image of the Muslim community
will be made clear in the following chapters.

From the axiomatic position achieved by the principle of serial revelation
it followed that that process should be described: temporally and spatially
fixed. The demand was met, indeed exceeded, by a supply of data formu-
lated to provide a meticulous chronology of the Quranic revelation.
Technically designated akhbdrjasbdb al-nuzul (reports about/causes of
revelation), this material found its most succinct expression in halakhic
exegesis, bearing eloquent witness to the many, often contradictory, uses
to which a single instance of revelation ,might be put. 2 The arbitrary
character of these data is best observed in the pseudo-historical literature
which accompanied, possibly preceded, the efforts of the halakhists to
distribute meaningfully the Quranic revelation over a period of twenty/
twenty-five years following the call of the Arabian prophet. The traditional
principle informing that literature, namely, that a chronology of revelation
is possible, has to my knowledge never been questioned. 3

A typical illustration is found in the accounts of an interview of Ja'far
b. Abi Talib with a/the ruler of Ethiopia (najdskt). The episode is one in a
narrative series treating of the Muslim emigration/exile to that country
(al-hijra al-ula) and of the dispatch by Quraysh of a mission to secure
return of the emigrants to Mecca. In the presence of those envoys Ja'far
was summoned to explain the circumstances of himself and his com-
panions. His response may be set out as follows.

1. We were a people, a folk in ignorance.

2. Worshipping idols.

3. And eating carrion.

4. Frequenting prostitutes.

1 Suyuti, Itqdn i, 82, 119, 121, 124.

3 GdQ i, 58-65; cf. below, IV pp. 126-7.

2 See below, IV pp. 177-8.

REVELATION AND CANON

39

5. And severing ties of kinship.

6. Violating the rules of security.

7. The powerful among us oppressing the weak.

Thus we were until :

1. God sent to us a messenger chosen from our midst, whose background,
honesty, faithfulness, and continence we knew.

2. Who called us to God, to associate no one with Him, to worship Him, to rid
ourselves of the stones and images which we and our fathers had been
worshipping. He commanded us to worship God alone and to associate
nothing with Him.

3. (And he commanded us) to abstain from that which is forbidden and from
blood.

4. And he forbade us prostitutes.

5. (And he commanded us) to observe the ties of kinship.

6. And to abide by the rules of security.

7. (And he forbade us) to devour the property of orphans.

8. And he commanded us to speak honestly and to act in good faith ; and (to
abstain from) false speech and the slander of honourable women.

9. And he commanded us to fulfil the duties of prayer, almsgiving, and fasting.

Recounted on the authority of an eye-witness (scil. Umm Salama) to the
event, this version was adduced in the earliest biography of the Arabian
prophet. 1

The substance of Ja'far's recital, designated by the tradent 'the principles
of Islam' {umur al-islam), is not unlike prescriptions of essentials for the
new faith published from time to time by the Christian apostles (e.g. Acts
1 5 : 20, 28-9). Formally, a degree of dramatic tension is achieved by the
syntactical balance developed round the phrase 'Thus we were until'. The
counterpoint consists almost entirely of locutions attested at least once, in
some cases frequently, in the canonical text of the Quranic revelation:

1. 2UutU JaI L,j3 \1S~ may be thought to reflect j^sj Ly ^Ijl

(Q. 11 : 29, 46: 23, cf. 7: 138, 27: 55), or even one of the four occurrences
of jahiliyya (interpreted as a collective: Q. 3: 154, 5: 50, 33: 33, 48: 26).
lL> UJI JUKI d*>o j^. is of common occurrence, e.g. J t^JUl y>
, V_^«j ^jZZ VI (Q. 62: 2, cf. 3 : 164, 9: 128). Of the four qualities attri-
buted to the prophet: nasab, sidq, amdna, and 'afaf, only the last is not
attested in scripture, though the remaining three nowhere appear as a
collective designation of prophetical credentials (cf. 4: 58).

2. |» LysVI Ju*j is found in Q. 14 : 35 and 26 : 71 , invariably related to the

1 Ibn Hisham, Sira i, 336; to underline the contrapuntal style of the oration I have
rearranged just slightly the components of the second series, which includes two items
not listed in the first. Variant and parallel traditions are found elsewhere, see Caetani,
Annali i, 866-7.

4 o QURANIC STUDIES

story of Abraham (cf. 6: 74 and the imagery of 19: 42-9). The somewhat
awkward statement l ya Ju*i L ^Ucij «-V"J e^*^ tJI UUoi
jUj'iflj SjUcsJI ^ dJji y UjUfj is such that it could provoke a reply
like Q. 1 1 : 62, but neither khalaa (rid) nor femora (stones) is so used in the
Qur'an. Awthdn (images, idols) is attested (Q. 29: 17, 25), but not a finite
form of wahhada. More easily placed is the paraphrase All I Ju*j ol U^lj
It-i iij^J V «JUj, reflecting Q. 7: 70 and 3: 64.

3. The phrase i^J I JStjj belongs to the series of prohibitions articulated
in Q. 2: 173, 5: 3, 6: 145, 16: 115, but for the complement in the contra-
puntal scheme neither kaff (abstention) nor maharim (prohibitions) is
Quranic, and the plural dimd' is not used of sacrificial blood.

4. ^^sJjaII jtij is frequently attested in the singular (e.g. Q. 4: 15, 19,
25) and the prohibition occurs (with harrama) in the plural in Q. 7: 33.

5. j»U.j VI £^juj appears as an admonition in Q. 47: 22, but the locution
p*. ji\ il^> is not scriptural (cf. however Q. 4: 1).

6. The tei-m jlja. is itself not found in scripture, but the exhortation
here reflects the contents of Q. 4: 36, and finite forms of the verb are
employed metaphorically ( !) in 23 : 88.

7. lL tSjill J^t>j is not scriptural, but its complement JL JSl
pjjuJI appears verbatim (plural) in Q. 4: 10, and varied slightly in 6: 152
and 17: 34.

8. This and the following injunctions are not adduced in Ja'far's
description of the Meccans prior to the appearance of the prophet, and
thus intrude upon the contrapuntal scheme of items (1) to (7). Their
phraseology is, however, Quranic: ^JbJI (JjJ^ may be related to <jLJ
tjJw» Q. 19: 50; J33JI Jjl occurs in 22: 30 (with awthdn, cf. 58: 2); ijju
oLyajcJI reflects c>\^a^J d_y>ji ^jJI m 2 4 : 4> 2 3 (with rama; qadhafa
is not so used in scripture).

9. ils^Jlj i'tA^aJI appear almost invariably together (e.g. Q. 2:43, 19:31,
55), and |»L^ separately (e.g. 2: 183) or with related injunctions (e.g. 2:

196 with hajj, sadaqa, and nusuk), but the combination of fasting with
prayer and almsgiving is not attested in scripture.

Now, the exact relationship between this very concise catechism and the
canonical text of the Quranic revelation is not immediately clear. Accep-
tance of the historicity of Ja'far's interview with the Najashi must lead one
to suppose either that the injunctions here expressed had been the subject
of revelations before the emigration to Ethiopia, or that they represent
prophetical logia later confirmed by or incorporated into the text of scrip-

REVELATION AND CANON

41

ture. On the other hand, the structure of the report suggests a careful
rhetorical formulation of Quranic material generally supposed to have been
revealed after the date of that event. Three further points in the interview
deserve notice. 1 Asked by the ruler to recite something from the revelation
sent to their prophet, Ja'far produced the beginning of Surat Maryam

( a*^ l S' ^ ljJL*>), at which the Najashi, greatly moved, declared: This is,
indeed, from the same source as that which Jesus uttered (il£jL» ^
S-kJj). Interrogated next day on his attitude towards Jesus, Ja'far replied:
He is the servant of God, His messenger, and His spirit, His word which
He bestowed upon the unblemished virgin Mary (<0^«jj «UJ| Ju* j&>
JjxJI JjAjJl pjj* ,J| Iftlilf «cul5"j JUjjj). Explicit mention of the nine-
teenth sura, possible allusion to the imagery of Q. 24: 35, and almost
certain reference to 4: 171-2 (but cf. also 19: 17, 30) might be thought to
strengthen the inference that the author of this report was familiar with a
fairly extensive range of Quranic diction. The positivist Caetani, failing to
recognize the literary form and reacting to what he regarded as anachron-
ism, dismissed the account in the Sira in favour of the terse version in
Tabari of the arrival of a delegation from Quraysh, in which appears no
reference to a conversation between Ja'far and the Najashi. 2 Caetani's re-
jection reflects his approval of a chronology of revelation in which Surat
Maryam, or at least its beginning, was Meccan, but in which much of the
content of Ja'far's confession of faith included material agreed to have
been revealed in Medina. The quite arbitrary character of that chronology
is clear from even a cursory examination of Muslim scholarship. In his
commentary to the Sira, Suhayli regarded the account of Ja'far and the
Najashi as unexceptionable. 3 Qummi stated explicitly Surat Maryam, not
merely its beginning, and adduced the Muslim catechism in a form almost
identical to that of an Apostolic promulgation. 4 Ibn al-Athir included an
account varying only slightly from that of the Sira, 5 And Suyuti cited the
episode approvingly. 6 It can, of course, be argued that reports (akhbar)
about and occasions/causes (asbab) of revelation are not quite the same
thing, and thus, that halakhic exegesis may disregard as Active, or at least
irrelevant, the data found in haggadic literature of the kind exhibited in
the Sira. But in practice such was seldom the case, and the foregoing analysis
may be thought to have suggested the interdependence of source materials
traditionally adduced as evidence of independent corroboration.

1 Sira i, 336-7.

2 Annali i, 277-8, mentioning only Q. 4: 171-2; Tabari, Amahs I/1189; similarly
GdQ i, 130; cf. Buhl, 'Beitrage', 13-22.

3 Al-Rawd al-unuf i, 213.

* Tajsir i, 176-9, ad Q. 5: 82.

5 Al-Kamil ii, 60-3. 6 Itqdn i, 50.

42

QURANIC STUDIES

I have proposed that one might interpret Ja'far's recital as a report of
prophetical logia exhibiting a stage of transmission prior to their incorpora-
tion into the Quranic canon. An essential feature of that report, and one
characteristic of the Sira as a type of exegetical literature, is the narratio.
This device, common in Biblical literature, provides a context of historical
circumstance or, at least, vaguely historical allusion, into which or by
means of which reports of a prophet's deeds (!TWS) or words (CTUI) may
be introduced. For Hebrew scripture the priority in time of such reports
over the actual reproduction in literary form of prophetical utterances
has been established. 1 To postulate a similar, if not identical, process for
Muslim scripture seems to me not unjustified, though in this particular
instance complicated by the redaction history of the Sira itself. 2 The fairly
consistent application of a contrapuntal scheme in Ja'far's confession
reveals, indeed, a preoccupation with rhetorical niceties, but certainly does
not preclude the possibility of oral transmission or delivery. Use of
counterpoint might even be interpreted as a mnemonic device calculated
to assist such. As set out in this episode the promulgation of basic Muslim
doctrine has the inestimable advantage of clarity and cohesion over its dis-
jointed and dispersed occurrence in the canonical text of scripture. That
difference may, at least superficially, indicate a composition posterior to the
several revelations contained there. From the same and similar evidence,
however, a rather more complex relationship between the canonical
traditions is discernible. In the narrative structure of the latter, revelations
of exclusively regulative content are presented in a manner which con-
forms to that characteristic of the Quranic theodicy as a whole, of which
the retribution pericopes (e.g. Shu'ayb) are examples.

In the story of Ja'far and the NajashI three distinct but related themes
may be seen : exile from the homeland, enumeration of the basic ingredients
of monotheism, and external recognition of the prophet's credentials. In
that way three purely literary functions — typological, exegetical, and
aetiological- — combine to produce a narrative history. The same process
can be observed in treatment of passages whose content is not regulative
but paraenetic: for example, in the relation of Sura 105 to the Affair of the
Elephant. 3 Whatever the sense, or variant readings, generated by the
phrase ashab al-fil, the literary tradition represents the elaboration of
three themes: holy war (Abraha's campaign), the inviolate sanctuary and
its protectors (Mecca and Quraysh), and the action of God in history (birds
as bearers of plague). By means of these, the elliptical Quranic passage
was incorporated into salvation history as an episode both lively and easily
understood. Here too, as in the story of Ja'far, relationship to the canon
is not merely causal, and an attempt to adduce it as illustrative of the

1 See von Rad, Theology ii, 34-9.

2 See below, IV pp. 127, 140-3.

3 Ibn Hisham, Sira i, 43-61.

REVELATION AND CANON

43

chronology of revelation was rejected by Suyuti. 1 Both episodes are charac-
terized by action which is symbolic and which takes precedence over the
direct speech of the corresponding canonical passages. The contrast
between parallel formulations, canonical and non-canonical, may be
likened to that which obtains between the narrative framework {Fremd-
bericht) and the dialogue (Ich-Bericht) observed in the retribution peri-
copes.

In all of this material, whether or not ultimately incorporated into the
canon, the organizing principle was the same: movement in history as an
act of God. From the foregoing analyses it may be agreed that the patterns
upon which description of that movement drew were ancient and well
established. Their literary formulation might seem to make superfluous
the question of historicity. And it is not merely the chronology of revela-
tion, as a recognized component of the Quranic masorah, which is proble-
matic. Traditional discussions of the modes of revelation would appear to
have derived considerable impetus from polemic about the precise distinc-
tion between the Muslim and Hebrew scriptures. Indeed, the history of
the canon itself provokes the question at least of methodological, if not
substantial, assimilation of earlier descriptive techniques. Assimilation of
the sort encountered here need not, of course, be defined in terms of a
rigorous distinction between Urerlebnis and Bildungserlebnis. A literary
analysis can, after all, only reveal what seems to be the essential role of
historiography, namely, the unceasing reinterpretation of tradition. It
need hardly be added that both method and material are almost infinitely
transferable.

• • ■ • *

In the exegetical literature formation of the Quranic canon is ascribed
to one of two processes: either official recognition of a corpus left intact by
the Arabian prophet (Urtext), or imposition of a uniform recension pro-
duced by an officially constituted body with concomitant suppression of
earlier and variant versions ('Uthmanic codex). The two processes may
appear as separate, if not quite mutually exclusive, traditions, or together
as two stages in a single tradition, the consequence of more or less successful
harmonization. Critical analysis of the tradition(s) is set out in the second
part of the fundamental work of Noldeke-Schwally, in which may be read
the authors' concluding judgement on the diametrical opposition between
formation of the Quranic canon and that of the Jewish and Christian
scriptures: 'Die Entstehung des muhammedanischen Kanons ist vollig
abweichend, man konnte sagen entgegengesetzt verlaufen. Er ist nicht das
Werk mehrerer Schriftsteller, sondern eines einzigen Mannes und deshalb
in der kurzen Spanne eines Menschenalters zustande gekommen'. 2 Now,

1 Itqdn i, 90; cf.> Horovitz, Untersuchungen, 10-11, 96—8; Blachfcre, Histoire iii, 788
n. 3. 2 GdQ ii, 120.

44

QURANIC STUDIES

it seems to me at least arguable that the evidence of the Qur'an itself, quite
apart from that of the exegetical tradition, lends little support to that
assertion. It may, indeed, appear from my description of that document
that the Muslim scripture is not only composite, but also, and such can
be inferred from a typological analysis of Quranic exegesis, that the period
required for its achievement was rather more than a single generation. 1

Traditional material relating to the canon must be assessed not merely
in respect of its intrinsic merit, but equally by reference to a series of data
extrinsic to the process itself of canonization but relevant none the less to
canonicity. It would seem not unreasonable to assume that the latter pre-
supposed acceptance by the Muslim community of the Quranic revelation
as normative for its religious life. And yet, Schacht's studies of the early
development of legal doctrine within the community demonstrate that,
with very few exceptions, Muslim jurisprudence was not derived from the
contents of tht„Qur'an. z It may be added that those few exceptions are
themselves hardly evidence for existence of the canon, and further observed
that even where doctrine was alleged to draw upon scripture, such is not
necessarily proof of the earlier existence of the scriptural source. 3 Deriva-
tion of law from scripture (halakhic exegesis) was a phenomenon of the
third/ninth century, and while the obvious inference is admittedly an
argumentum e silentio, the chronology of the source material demands that
it be mentioned. A similar kind of negative evidence is absence of any
reference to the Qur'an in the Fiqh Akbar I, dated by Wensinck to about
the middle of the second/eighth century. 4 Moreover, stabilization of the
text of scripture (masoretic exegesis) was an activity whose literary ex-
pression is also not attested before the third/ninth century, and the
appearance of the classical masdhif literature (variae lectiones) was even
later. 5 It is of course neither possible, nor necessary, to maintain that the
material of the canon did not, in some form, exist prior to that period of
intensive literary activity, but establishment of a standard text such as is
implied by the 'Uthmanic recension traditions can hardly have been earlier.

One has further to consider the actual significance of the textual variants
exhibited in these traditions. From the material assembled by Bergstrasser
and Pretzl, and in more detail by Jeffery, it could well be asked to what
extent any of the variants, or variant codices ( ?), may be said to represent
traditions genuinely independent of the 'Uthmanic recension. 6 The
infinitesimal differences are not such as would seem to have necessitated
suppression of the non-'Uthmanic versions, the more so since a minimal
standard deviation from the canon was accommodated by the several inter-

1 See below, IV pp. 145—8.

1 Cf. Strack, Introduction, 10.

5 See Jeffery, Materials, 1-16.

6 GdQ Hi, 57-115; Materials, 19-355-

2 Origins, passim but esp. 284-7.

4 Creed, 102-24.

REVELATION AND CANON

45

pretations of the ahruf doctrine. 1 The choice between interpreting those
codices as conscious (i.e. exegetical) variations upon the 'Uthmanic
recension, or as having in common with that recension an earlier Vorlage,
is not an easy one. What might be considered a special category of variant
codex, the 'metropolitan codices' (masahif al-amsdr), do not display the
differences either among themselves or from the 'Uthmanic recension which
are alleged to have provoked the editorial measures attributed to the third
caliph. 2 The tradition itself of separate collections of amsar-variants, far
from being attributable to so early an authority as the Damascene muqri'
Ibn 'Amir (d. 1 18/736), appears not to be more ancient than Farra' (d. 207/
822) or possibly than his teacher Kisa'I (d. 189/804). Fixed terminology,
like reference to the 'Uthmanic recension as imam, or to mushaf as codex,
in contrast to the plural masahif not as codices but as variants, can also not
be dated earlier than the beginning of the third/ninth century. 3 The ran-
dom predominance of Medinese readings in both Sibawayh and Farra',
where one might have expected reflections of practice in Basra and Kufa,
cannot but provoke the impression that concern with the text of scripture
did not precede by much the appearance of the masoretic literature as it
has in fact been preserved. Failure to situate regionally actual manuscripts
of the Qur'an by collating them with collections of 'metropolitan' variants
might be thought to confirm that impression. 4

Now, confronted by the 'Uthmanic recension traditions one is compelled
to assume either that suppression of substantial deviations was so instantly
and universally successful that no trace of serious opposition remained,
or that the story was a fiction designed to serve another purpose. The first
possibility is, however, belied by the chronological sequence of literature
on the Qur'an, of which none may be described as presupposing a standard
or ne varietur text as early as the middle of the first/seventh century, and
further, by absence of explicit reference to a canon in contexts where such
ought to appear. The second possibility, on the other hand, gives rise to
some interesting but naturally inconclusive speculation on the means to
which a newly independent religious community might have recourse in
the effort to describe its origins. I have alluded to the likelihood of a
Rabbinic model for the account of an authoritative text produced in com-
mittee, namely the Jamnia tradition on the canonization of Hebrew scrip-
ture. 5 Similarly, the Muslim tradition of an Urtext, whether conceived of
as independent or as a stage preliminary to 'Uthman's editorial work,
might be thought to reflect Rabbinic views on the Mosaic reception of
the Torah. That conjecture derives some support from the relationship

' GdQ iii, 186-90, 313-28.

2 Cf. the lists in Ibn Abl Dawiid, apud Jeffery, Materials, 39-49; and GdQ iii, 6-19.

3 GdQ iii, 9 n. 4; Jeffery, op. cit. 13; Beck, 'Kodizesvarianten', 353-76.

4 GdQ iii, 270: 'dass fast alle Kodizes einen Mischtext aufweisen*.

5 See above, p. 21.

46

QURANIC STUDIES

between Moses and Muhammad as recipients of the word of God, integral
to Muslim discussion of the modes of revelation. The primary source of
strength for the Urtext tradition may be found in the concept of 'the final

review' (S^^l i^jJI), representing the culmination of a series of en-
counters between the prophet and the angel Gabriel for periodic organiza-
tion of the material so far revealed 'in order that abrogated matter might
be distinguished from that which remained effective' (^J U ^> ^\
Jii Uj) or 'abrogated distinguished from its replacement' (J Jb Lj U). 1
The significant element in these descriptions is reference to the doctrine
of abrogation (naskh) and hence the implications of the Urtext theory
for halakhic exegesis. 2 Later formulations of that doctrine required that
the process of abrogation be incomplete upon the death of the prophet
and concomitantly that definitive organization of the text of revelation be
postponed until after that date. The 'Uthmanic recension story-may be
regarded as the means by which that end was attained. 3

Thus could the two canon traditions be seen as complementary rather
than contradictory, though the actual instrument of harmonization was the
celebrated codex of Ilafsa, which provided a tidy sequence of events cover-
ing the period from the death of the prophet to the action of the caliph
'Uthman. 4 It is, however, not absolutely necessary to select only one of two
interpretations: the 'Uthmanic recension story as a reflex of the Rabbinic
academy at Jamnia, or as a logical construction of the halakhists. The
technical term employed to describe 'Uthman 's editorial work, scil. jam,
was used with a semantic latitude capable of accommodating a number of
related but quite distinct actions. Suyutl's synthesis includes all of them : in
naw' 18 he assumed throughout the equivalence jam' : tartib (arrangement),
but distinguished arrangement/collection between two covers (i.e. in codex
form), internal arrangement of suras, and arrangement of/restriction to
readings confirmed by the authority of the prophet, and insisted as well
upon the difference between order of revelation {tartib al-nuzul) and order
of recitation [tartib al-tildwa). s In naw' 20, traditions were adduced in
which jam' could be interpreted as preservation/memorization (hifz), as
recording by writing (kitdba), or as hearing and obeying (al-sam' wal-ta'a
lahu). 6 Such a spectrum of meaning, besides reflecting a series of doctrinal
positions in that long discussion, attests surely to uncertainty about the
process by which revelation became canon.

I have in the preceding pages attempted to show that the structure itself

1 Suyuti, Itqan i, 142, and Kitab al-Mabdni, 26, respectively.

2 See below, IV pp. 192-201.

3 Cf. Burton, 'Cranes', 246-65, esp. 260; id. 'Collection', 42-60.
* GdQ ii, zo-3, 91, 114-15.

s Itqan i, 164-83; cf. also Kitab al-Mabdni, ch. Ill, 39-77.
6 Itqan i, 199-206.

REVELATION AND CANON

47

of Muslim scripture lends little support to the theory of a deliberate
edition. Particularly in the exempla of salvation history, characterized by
variant traditions, but also in passages of exclusively paraenetic or eschato-
logical content, ellipsis and repetition are such as to suggest not the care-
fully executed project of one or of many men, but rather the product of an
organic development from originally independent traditions during a long
period of transmission. That such traditions might have been of local/
regional character is not impossible, but in view of the inconclusive nature
of the so-called 'metropolitan codices' regional distribution of the variant
traditions could hardly be justified. An alternative and less refractory hypo-
thesis is one already advanced: juxtaposition of independent pericopes to
some extent unified by means of a limited number of rhetorical conventions.
Such might be held to account both for the repetitive character of the
document and for what is undeniably its stylistic homogeneity, the latter
quality in part a consequence of the former. 1 The content of these pericopes
may be described as prophetical login whose formulation exhibits a number
of recognizable literary types based on what I have designated schemata
of revelation. In canonical form these logia are expressed, not quite con-
sistently, as the direct utterance of God, but outside the canon take the
form of reports about such utterances. An example of the latter was seen
in the story of Ja'far b. Abi Talib and the Najashi.

Whether one is justified in equating non-canonical with pre-canonical
is a historical problem complicated rather than clarified by the evolution
of exegetical literature. Of the earliest form of that literature, described
below as narrative/haggadic, we have no specimens which do not exhibit
traces of redaction characteristic of the third/ninth century. Thus the forms
in which prophetical logia were likely to have been transmitted display a
version of scripture which is unmistakably canonical. Primacy of the
narratio is none the less evident, and stylistic as well as explicative ele-
ments indicate oral transmission. 2 Again, Ja'far's recital is instructive, for
there material of an ethical and regulative nature was presented in a form
both hortatory and entertaining, but above all, owing to the contrapuntal
scheme employed, in a manner both easy to understand and to remember.
The use of symmetry as a mnemonic technique in oral transmission is of
course widely attested. 3 Repeated application of conventional formulae
of introduction and conclusion, refrains, litanies, and the structural balance
observable in related pericopes perform an identical service. It might, on
the other hand, be more accurate to speak not of oral transmission but of
oral delivery. Resort to mnemonic device and symmetrical structure does
not preclude the existence of written Vorlagen, indeed, in many instances

1 See below, III pp. 1 17-18.

2 See below, IV pp. 129-31, 145-8.

3 C£ Gerhardsson, Memory, 147, but see the entire section 136-56.

48

QURANIC STUDIES

presupposes such. Coexistence of textual transmission and oral tradition
may be substantiated not only by the technique of dictation, but also by
reference to the exigencies of typically cultic situations (liturgical and didac-
tic), in which the more appropriate oral delivery exhibits a refinement of
simple and straightforward recitation from memorized texts. That the many
techniques associated with refined and sophisticated oral delivery could
ultimately be incorporated into an improved text may be cited as evidence
of rhetorical development, but not of change from exclusively oral to
exclusively written transmission. 1

Related to procedures of transmission and delivery are the techniques
of oral and written composition. Though from the point of view of his-
torical reconstruction it is unquestionably useful, perhaps imperative, to
keep separate the discussion of each, mention at least of the problems
relating to composition is not unjustified. Both the very high fre-
quency and the uniform distribution in the Qur'an of formulae and of
'formulaic systems' could indicate not only a long period of oral trans-
mission but also of oral composition. 2 Analysis of formulaic language in
Muslim scripture would include statistics for the thematic permutations
which I have described as variant traditions, as well as for the schemata
adduced to illustrate the theodicy. Equally important are the rhyme-
phrases employed in Quranic periodization, which exhibit, in addition to
the stressed syllable of the rhyme itself, fairly uniform length and thus a
nearly constant metrical value. 3 Those phrases are most conspicuous in
passages of halakhic and narrative content, where they serve as both con-
junctive and disjunctive markers. 4 Now, all of that material — theme,
schemata, and rhyme-phrase — may be described as components of formu-
laic systems, but not necessarily as proof of oral composition. The imagery
and lexicon of Muslim scripture are almost exclusively archetypal and
suggest, if they do not presuppose, some contact with literary precursors.
The dichotomy postulated between 'borrowing' and 'traditional language'
is possibly misleading and certainly an oversimplification: like most
linguistic expression the structure of monotheist revelation contains very
little that is not 'traditional language'. 5 For the Quranic revelation ascrip-
tion to Biblical archetypes has been, perhaps unnecessarily, complicated
by the existence of a native Arabic tradition of monotheistic (kamft) poetry. 6
The authenticity of that poetry has been disputed; its importance to the
Muslim exegetical tradition cannot be. But the sources of archetypal

1 Gerhardsson, loc. cit. ; cf. Widengren, 'Oral tradition', esp. 201-33.

2 Cf . Culley, Formulaic language, 1 0-20, 2 1 -7 ; Muilenburg, ' Hebrew rhetoric', 97- 1 1 1 .

3 The criteria for detection and assessment of formulaic language were after all derived
from oral poetry, see references in Culley, loc. cit.; and Monroe, 'Oral composition',
1-53-

4 See below, Hipp. 112, 115-6.

5 Cf. Culley, op. cit. 112-19. 6 See below, III pp. 96-7.

REVELATION AND CANON

49

imagery in the Qur'an are not thereby concealed, nor is the sheer quantity
of reference to Patriarchal narrative in any way diminished (Moses: 502
verses in 36 suras, Abraham: 245 verses in 25 suras, Noah: 131 verses
in 28 suras). 1 Elaboration of the corpus of prophetical logia from which
the Quranic canon was eventually separated may have been essentially
a product of oral composition. Emergence of the canon itself, however,
represented application of considerable literary technique. Not the least
of the problems provoked by its final form is the erratic distribution
of obviously related pericopes. 2

Analysis of its parts does not thus necessarily explain existence of the
whole. Lack of such over-all logical structure in the Qur'an as is found in
the Jewish and Christian canons, is reflected in traditions which attempt
to postulate a correspondence (vaguely defined and somewhat contrived)

between parts of the Muslim canon and the earlier scriptures, e.g. Jli

J-vsuUU c-lyLaj JJij^\ ot£«- 3 The descriptive terms are merely quantita-
tive and the correspondence quite arbitrary, but that it should have been
adduced at all is worthy of remark.

The fact of canonicity may be seen as a kind of watershed in the trans-
mission history of the Quranic revelation. 4 Development beyond that point,
which I should hesitate to set before the end of the second/eighth century, is
to be elicited from a study of exegesis and commentary. Description of the
course of events up to that date is, I have more than once suggested,
frustrated by the form in which pertinent witness has been preserved. Any
attempt at reconstruction is thus hazardous, being limited to tenuous con-
clusions from literary analogies. If the pericope hypothesis is acknowledged
to make some sense of the Quranic data themselves, it requires none the
less to be supplemented by a notion of the environment in which pro-
phetical logia might have been preserved and transmitted. Now, what could
be seen as obviously analogous circumstances, namely those obtaining for
preservation and transmission of both Rabbinic and Apostolic formulations
of the word of God, presuppose for both an authoritative centre of such
activity, which was Jerusalem. 5 Despite implicit emphasis upon the role of
Medina in the 'Uthmanic recension traditions, evidence for a single centre
of activity is not easily found in the pre-canonical transmission history of the

1 Moubarac, Abraham, 27-9.

1 Cf. GdQ ii, 63-8 : with regard to 64 n. 1 it may be observed that Geiger's proposal
that the Mishnah tractates were arranged in descending order of their length was ulti-
mately rejected by Strack, who suggested that their sequence corresponded to that of
legal problems in the Pentateuch, see Introduction, 27-8.

3 Kitdb al-Mabdni, 235 ; variants and discussion in Suyuti, Itqan i, 163, 177-81.

4 Cf. Koch, Growth, 106-8; and Vermes, Scripture, esp. 127-77.

5 Cf. Gerhardsson, Memory, 214-20.

5°

QURANIC STUDIES

Qur'an, nor, for that matter, in the early development of Islamic juris-
prudence. 1 If the origins of exegetical literature may in fact be located in
Mesopotamia, that is the paradoxical corollary of a social and political
development, the literary description of which was designed precisely to
demonstrate that the story of Islam was conterminous with the history of
the Arabian peninsula, and especially of the Hijaz in the first/seventh
century. As such the descriptive material need not be discounted, but can
by no means be accepted as constituting disinterested historiography. Thus
the fact of the Mesopotamian environment emerges, perhaps not quite
accidentally, and must be adduced as criterion for assessing any evidence
which purports to describe the circumstances of Islam prior to the third/
ninth century. 2

Enough has been said of the canonization traditions to indicate their
contradictory, and probably polemical, character. Proposed as alternative
was the concept of an organic development exhibiting gradual juxtaposition
of originally separate collections of logia. The failure to eliminate repetition
in the canon might be attributed to the status which these logia had already
achieved in the several ( !) communities within which they originated and
by whose members they were transmitted. Here 'community' need not be
understood as a regional specification, though such is not impossible.
I should be inclined to postulate the growth of logia collections in environ-
ments essentially sectarian but within the mainstream of oriental mono-
theism. Such an environment could be inferred from the evidence of
parallels proposed by Rabin between Islamic terminology and that of the
Qumran sect. 3 But some of that material reflects polygenesis rather than
diffusion through historical contact, e.g. the light imagery and the per-
sonification of evil. 4 And resemblance can be deceptive : rather than in the
cognate fassara I am tempted to see a methodological reflex of Qumranic
pesher in the Muslim term tawil, and thus an 'inverted' semantic relation-
ship between pesher and tafsir, similar to that found by Rabin to obtain
between Qumranic and Rabbinic terminology. 5 The primary difficulty,
however, in this and all such expositions aimed at demonstrating historical
diffusion lies in their uncritical assent to the traditional chronology of
Islamic origins, resulting inevitably in socio-psychological analyses of data
whose literary, rather than historical, character is patent. 6 Some scholars,
among them Ben-Zvi and Katsch, have been excessively generous in their
assessment of the documentary value of Islamic source materials for the
existence and cultural significance ( !) of Jewish communities in the Hijaz,

1 See Schacht, Origins, 8, 213.

2 On geographical factors in the assessment of legal source materials, cf. Schacht,
Origins, 188-9, 22 3. Z2 8. 3 Qumran, 113-30.

4 Qumran, 114-15 and 122, resp.

5 Qumran, 96-7, 1 08-11, 117; cf. below, IV p. 246.

6 e.g. Rabin, Qumran, 120-7; cf. my observations in BSOAS xxxiii (1970) 613-14.

REVELATION AND CANON

51

about which Jewish sources are themselves silent. 1 References in Rabbinic
literature to Arabia are of remarkably little worth for purposes of historical
reconstruction, and especially for the Hijaz in the sixth and seventh cen-
turies. 2 The incompatibility of Islamic and Jewish sources was only
partially neutralized, but the tyranny of the 'Hijazi origins of Islam' fully
demonstrated, by insistence upon a major Jewish immigration into central
Arabia. 3 Some of the material assembled by Rabin, such as apocalyptic
concepts and embellishments to prophetology, represent of course diffusion
through contact, but do not require an exodus from Judaea into the Arabian
desert. 4

Development and perpetuation of a logia tradition by the sectarian
group/community, wherever it (they) may have been situated, can be
ascribed to the exigencies of cult or of instruction, but probably not to the
requirements of legislative or judicial authority. The latter must be
regarded as the agent itself of canonization and posterior to the liturgical
and didactic functions of the tradition. The entire process of canonization
will thus be seen as a protracted one of community formation (Gemeinde-
bildung). 5 The essentially cultic/didactic role of the logia tradition is
explicit even in the term qur'dn {lectio, legenda), which may be said to form
within the lexicon of the canonization traditions a kind of binary opposition
with the term mushaf (codex). 6 In that context the reproach levelled at the

caliph 'Uthman is instructive: ^1 \g5~ j& Ld^oT/JI jlS"- 7 The caliph's
reply that that the Qur'an (sic) was in fact one and had been sent from One,
exhibiting a post-canonical stage of the discussion, cannot delete the im-
pression that qur'dn originally designated any one of several logia collec-
tions. Such is of course confirmed by the masahif literature. Canonicity
once achieved, qur'dn and mushaf became synonymous as designations of
revelation. That function had, however, to be shared with a third term:
sunna (exemplum), in which was symbolized the definitive enthronement of
revelation as canon for the Islamic community. The act found succinct
expression in SuyutI, in his observation on a typology of revelation (kaldm

alldh) articulated by Juwayni: y* Jj'ill ^iJlj JliJI y> jT^JI cii
^'jji^ Cr 4 -? Ol iiyi 1*5"*^*J1j J)~! tils' iMj^ 'j' L^iuJI

1 Ben Zvi, 'Les Origines', esp. 178-90; Katsch, Judaism, xxv.

2 Cf. Krauss, 'Nachrichten', 321-53; Cohen, 'Arabisms', 221-33; Steinschneider,
Polemische Literatur, Anhang VII, esp. 244-73; Hirschberg, Lehren, 14-26, but see
below, III pp. 96-7.

3 See especially Torrey, Foundation, 1-27; but also Obermann, 'Islamic Origins',
58-120.

* Qumran, 11 8-21, 128; cf. Wieder, Scrolls, 4: on the messianic symbolism of 'wilder-
ness'.

5 Cf. Koch, Growth, 44; Seeligmann, 'Midraschexegese', 150-81.

6 Horovitz, Untersuchungen, 74.

7 Tabarl, Annates I/2952; cf. GdQ ii, 50, 90.

S 2 QURANIC STUDIES

iaUJb. 1 The distinction between permitted modes of transmission {Utter atim
and paraphrastic) was merely formal recognition of a principle which
remained almost purely theoretical. In practice Sunna, as revelation,
was transmitted with infinite care, and was the primary instrument by
which the Quranic revelation was linked with the historical figure of an
Arabian prophet. Inherent in the reports pertaining to that process are two
factors worthy of remark: the immediacy and practicability of ins con-
suetudinis as articulated in the Sunna; and consistent emphasis upon the
role of Medina as its source and paradigm. 2

Formulation of the Sunna as embodiment of prophetical practice/
judgement cannot be dated before the beginning of the third/ninth century,
and thus may be seen as coincident with recognition of the Qur'an as the
canonical collection of prophetical logia. Juxtaposition of the two revela-
tions as equally authoritative need not, in my opinion, be understood to
imply that Qur'an yielded a position already secure to the encroachments
of Sunna. 3 It can indeed be argued that the opposite was so : that canoni2a-
tion of the Quranic revelation could only have been effected within the
community once its content could be related to that of the prophetical
Sunna and, perhaps more important, to the historical figure delineated
there. Acknowledgement of a prophet as source of regulative prescription
for the religious life of the community may be thought to reflect traditional
notions of charismatic authority. That such should entail discussion of
the emblems, and in particular the credentials, of prophethood cannot be
surprising. In Muslim literature a not inconsiderable portion of that
discussion was concerned to establish the role of scripture as testimony to
prophethood, and it is that function of the Quranic revelation which forms
the subject of the following chapter.

1 Itqdn i, iz8 ; and Goldziher, 'Kampfe', 86-98.
* Goldziher, Studien ii, 1 1-2Z.

3 Pace Goldziher, Studien ii, 20; see below, IV pp. 174-5, 176-7.

II

EMBLEMS OF PROPHETHOOD

A marked feature of Islamic prophetology is its ethnic orientation,
nowhere more clearly expressed than in the Quranic concept of umam
hhaliya. 1 But the scriptural imagery did not quite imply ethnic exclusive-
ness, and even an apparently unambiguous passage like Q. 14 : 4 ^ ULj ] Lj
<UjS jLL V I i}j~>j could be adduced in support either of the ethnocentric
nature of the prophet's mission (by stressing the basically Arabic content
of revelation) or as divine proof of its universality (by stressing the inclusion
of non- Arabic idiom). 2 Commentary on Q. 14 : 4 consists largely of specula-
tion about the language(s) of scripture and of God, a topic complicated by
controversy over the origins of classical Arabic. 3 Apart from the dia-
metrically opposed possibilities of that particular verse, it may be asserted
that Quranic imagery underlining the ethnocentric position of prophets

is both uniform and consistent, e.g. J ^ L»l JS3 _j (Q. 10 : 47), j U JSO j
(13: 8), jjjj I4J VI SJ & Ob (35:24), l^sfrA i-i JS" Uc^j (28: 75).
These aphoristic formulations, slightly modified, are reflected in the
retribution perieopes, especially in the conventions employed to introduce
the themes of commission and of rejection. For example, in Q. 16: 113

Ml

6jij& Jj^j j^U. JjaJj (cf. 23: 44, 30: 47, 38: 4, 50: 2, 71: 1) is

emphasized the prophet's membership of the community to which he has
been sent, as also in the designation akh (brother) employed in conjunction
with ahl (people) and ashab (members/companions) in the commission
formulae. 4 Both components, membership and rejection, belong of course
to the traditional imagery of prophetical experience, e.g. Deuteronomy 18:
18, Matthew 5 : 12, Luke 6: 23. Application of the same or similar descrip-
tions not only to B. Isra'il but to the entire range of umam hhaliya exhibits
the attenuated election tradition found in Rabbinic and Christian literature,
according to which every nation was recipient of a prophet, e.g. Numbers
Rabba 20: 0V1 sn nwixV tbsjstd bvrwb anoan casm ffoba *r»sra apa.s
That urnmot ha-olam, moreover, is reflected in Quranic ummiyyun (gentiles)

1 See above, I pp. 3-5. 2 Cf. SuyutI, Itqan ii, 106-7.

3 See below, p. 81, and III pp. 93-106.

4 See above, I pp. 24-5.

5 Speyer, Erzdhlungen, 417-18; cf. Andrae, Person, 292-3, adducing other parallels;
Horovitz, Untersuchungen, 46; Wensinck, 'Propheten', 185.

54

QURANIC STUDIES

can hardly be doubted, and not only in Q. 3 : 30, 75 and 62 : 2, but also
in 2: 78. 1 Reference to ummiyyun among Jews, as in Q. 2: 78, may of
course be a reflex of 'am ha-ares, but probably in the sense of exclusion or
separation, as in Ezra 10: 11 and sectarian (Pharisaic, Qumranic, Karaite)
applications, invariably abusive. 2 Whether, on the other hand, the locution
(Q. 7: 157-8) belongs properly to this instance of
linguistic and conceptual assimilation is a separate problem; its common
interpretation suggests a parallel to if not a caique of 'am ha-ares. 3

From the notion of a prophet for every nation the universal character
of each prophet's mission may be elicited, as was argued for Q. 14: 4,
or by reference to the epitheta ornantia lil-alamin (21: 107, 25: 1) and
'aid 'l-'alamin (3: 33, 6: 86). An internal relationship governing the dis-
patch and destiny of all prophets may be seen in the concept fatra (Q. 5 :
19), understood there to ensure that no nation be without a prophet.
A unique instance of transition from ethnocentric to universal mission is
explicit in scriptural references to the figure of Abraham, e.g. Q. 3 : 67 U
UJ....J LLj>» jj^-b Uil Vj L>-5_^> (»-*>I j k'I an d 10: 120 (jls" f*--*!^! jl
LLv«. <UJ Uiti The imagery is traditional, exhibited in the evolution
from God-seeker to patriarch to the Pauline concept 'father of all the faith-
ful' (e.g. Romans 4: 9-12). 4 For the epithet umma in Q. 16: 120 Zamakh-
sharl provided the gloss mamumjimdm (exemplum: cf. 2: 124), a conclusion
of methodological if not material interest: the obvious Vorlage in Genesis
12: 2 VnJ ""U 1 ? "jtPVKl was not thereby concealed. 5

Rigorous and consistent distinction between the designations nabi and
rasul is not justified by Quranic usage, though something is to be said for
linking the term rasul (apostle, messenger) with the concept of mission to
a specific nation (umma). 6 hike nadhir, mundhir, bashir, mubashshir, and
even 'abd, the denomination rasul is basically functional; the only generic
term for prophet is the Hebrew loan nabi. But while, or perhaps because,
Quranic nabi is used only of Biblical figures, the generic employed in
exegetical literature was mursal, a Quranic term (e.g. Q. 7: 75) understood

1 Pace Horovitz, op. cit. 53.

2 Cf. Rabin, Qumran, 12-18, 61-4; his reference 125 n. 2 to Q. 2: 78 must(?) be to that
use of 'am ha-ares; see Wieder, Scrolls, 153-6, and the references in Paret, Der Koran,
21-2 ad loc; from the context of Q. z: 78-9 inference of an allusion to sectarian strife
within the Jewish community seems justified: for the vituperative epithets in that kind
of polemic cf. Wieder, op. cit. 129-60.

3 See below, p. 63.

4 Gerhardsson, Memory, 287-8; Beck, 'Abraham*, 89-94; Moubarac, Abraham,
99-118, but also 140-8: stressing, curiously, Abraham's role as recipient of revelation;
cf. also Chapira, 'Legendes', 86-107, 37-43.

5 Kashshdf ii, 641-2 ad loc.; cf. Geiger, Was hat Mohammed, 26-8, 201; Torrey,
Foundation, 38; Katsch, Judaism, 75-6.

6 Cf. Wensinck, 'Propheten', 17 1-5; and the modifications proposed thereto by Horo-
vitz, Untersuchungen, 48-9.

EMBLEMS OF PROPHETHOOD

55

to include both nabi and rasul, as in Suyuti's enumeration of those men-
tioned in Muslim scripture. 1 Election to prophethood, a divine and uni-
lateral act, is regularly expressed by one of three verbs: ikhtdra (Q. 20: 13
Moses), ijtaba (6: 87 Ishmael and Hebrew prophets), and istafa (2: 130
Abraham), each employed collectively as well as individually. The generally
factitive verb ja'ala, with an appropriate objective complement, may per-
form the same function (e.g. Q. 19: 31, 26: 21), and an isolated instance of
island a (20: 41 Moses) may be so interpreted. That the passive participle
mukhlas is included in the election imagery, as proposed by Lidzbarski,
can be inferred from Q. 19: 51 (Moses), possibly from 12: 24 (Joseph), but
not in the sense of prophetical election for the several occurrences of the
plural (e.g. 37 : 40), in which the image is that of the community of worship-
pers. 2 The single occurrence of the participle mustafayn (Q. 38: 47 plural)
is, however, a reference to prophetical election. The very large and merely
symbolic numbers related of prophets in the Muslim tradition appear to
reflect discussion of the respective merits of angels and prophets, a relation-
ship derived from the semantic proximity of malak to rasul (cf. Malachi 3 : 1
'OX'W nVtP 'ttn) and crystallized in the messenger formulae. 3 In Q. 40: 78,
a passage commonly adduced in argument for an infinite number of divine
messengers, the plural rusul is employed, and glossed nabi in the exegetical
tradition. 4

Of the later doctrinal development which granted the Arabian prophet
superiority (sayyid al-mursalin) over God's other emissaries, there is no
unequivocal trace in Muslim scripture. 5 Indeed, such statements as Q. 2 :
285 aLj ^ .U.I ^ (Jj^iJ V (similarly 2: 136, 3 : 83) clearly make the opposite
point, namely, that among prophets there was no distinction in rank. That
view is also emphasized in Q. 41 : 43 dJJJi ^ J-^A) ^ Ji L tLU Jli> L
and 46: 9 J-^JI U-Jij c^S" L Ji in both of which the addressee was
traditionally seen to be Muhammad. In flat contradiction to such passages
is Q. 17: 55 jiujt i3~rr^ ^-2^ -^J ( c f- *h e imagery in 17: 21) and
by implication 4: 125 ^LU- p~*>\ ^jI <U)I A?=jU (identical phraseology differ-
ently applied in 17 : 73, 25 : 28), of which the latter may be compared with
•anx Dniax of Isaiah 41: 8. 6 Moreover, Q. 4: 171-2 and 19: 19 may he
thought to accord a special distinction to Jesus, while a number of verses

1 Itqan iv, 58-67; cf. Speyer, Ersahlungen, 416 n. 1.

2 'Salam und Islam', 95-6.

3 See above, I pp. 12-13; cf. Horovitz, op. cit. 46; Wensinck, op. cit. 184, and id.
Creed, 200-2, 204.

* e.g. Zamakhshari, Kashshdf iv, 180 ad loc.

5 See Andrae, Person, 245-89, esp. 247-50 for Sulami's systematic elaboration of that
argument.

6 Cf. Geiger, Was hat Mohammed, 119-20; Speyer, Erzdhlungen, 173.

56

QURANIC STUDIES

based on the imagery of Q. 2: 253 (those prophets to whom God spoke)
allude to Adam (2: 31, 37), Abraham (2: 124), but especially Moses
(4: 164, 7: 143, 26: 10, 27: 8-12, 28: 30-5), reflecting thus the data of the
exegetical tradition. 1 Such as it is, the scriptural material may be enlisted
to support the particular position of Moses in the prophetical hierarchy,
but hardly that of Muhammad. The paradigm was not only Biblical, but
Rabbinic. 2

The certainty that already in the Hexateuch the figure of Moses was the
product of literary elaboration is of some relevance to a description of the
analogous process for Muhammad. 3 There, however, the literary develop-
ment is confined to the non-canonical revelation : the prophetical Sunna
together with a Muhammadan evangelium formulated as a history of the
Hijazi Arabs. 4 Like its Mosaic Vorbildthe portrait of Muhammad emerged
gradually and in response to the needs of a religious community. 5 But
unlike the Hexateuch, from which could be inferred at least the outlines
of a historical portrait of Moses, the role of the Qur'an in the delineation
of an Arabian prophet was peripheral: evidence of a divine communi-
cation but not a report of its circumstances. The historical value of
Muslim scripture lies, it seems to me, not in its role as source for the
biography of Muhammad, but rather as source for the concepts eventually
applied to composition of the Muslim theology of prophethood. The latter
are both directly accessible in the text of scripture and susceptible of sche-
matic realization, while the very notion of biographical data in the Qur'an
depends upon exegetical principles derived from material external to the
canon. The satisfaction with a century of Quranic studies expressed by
Paret is thus in my opinion hardly justified. 6 His recognition of the
arbitrary value of variae lectiones and of the real contrast between ortho-
graphic and interpretative variants, as well as of the problems posed by
parallel passages, could have provoked some doubt about the reliability
of the 'Uthmanic recension traditions, rather than questions about
Muhammad's articulation of dogma. 7 Further, his distinction between an
'ausserer Geschichtsablauf ' and an 'innere Einstellung', and his admission
that only for the latter can the Qur'an be of some documentary value, is
merely a reflex of the now well-established method of psychological
description. 8

That the prophetical Sunna itself contains ample evidence of community
practice (ius consuetudinis: sunna muttaba'd) as well as practice ascribed

1 See above, I pp. 35-8; and cf. Torrey, Foundation, 75-82 vs. Wellhausen.

2 Cf. Speyer, op. cit. 410-20; Jeffery, 'Scripture', 108-9, 126-7; Katsch, Judaism,
172-3.

3 Von Rad, Theology i, 289-96. 4 See below, pp. 65-73.

5 Cf. Andrae, Person, 186.

6 'Der Koran als Geschichtsquelle', 24-42. 7 Paret, op. cit. 28, 31-2.
8 Paret, op. cit. 33 ff; see above, I pp. 43-4, 51-2.

EMBLEMS OF PROPHETHOOD

57

specifically to the prophet {sunned al-nabi) is clear from the well-known
document conveying the caliph 'Umar's instructions to the qadi Abu Musa
'1-Ash'ari, in some recensions of which sunna was transmitted (or glossed)
sunnat al-nabi. 1 I consider at least questionable Margoliouth's assumption
that the two sources of law were (a) texts of the Qur'an, and (b) practice. 2
From the document itself it is impossible to insist upon a neat distinction
between community and prophetical practice, or between 'texts of the
Qur'an' and what I have in the foregoing pages designated prophetical
logia. Qur'an was anyway glossed or transmitted in some recensions kitab
allah, which may not have been the same thing. 3 The purpose of the
caliph's letter was of course justification of resort to analogy, and the
authenticity of the document may well be doubted. 4 The contrast between
sunna as practice of the community and sunna as practice of the prophet,
analogous to the Talmudic distinction between minhag and halakhah, 5
could be and was neutralized by recourse to a simple and transparent
expedient: elevation of all statutes, whatever their origin, to the status of
revelation vouchsafed to a single identifiable recipient. A Muslim formula-
tion of that dogma reads OTyJI c~jjI jf pL,j <*-!& <Ul! JUj
2LuJI <*j^. 6 Recognition of the (prophetical) Sunna as Mishnah may

be regarded as yet another element in what could be described as the
'Mosaic syndrome' of Muslim prophetology. Within the community the

didactic principle of imitatio magistri (JjcSVl J^l) was realized as
magister dixit (cf. "^DS iWKh> nsVn), in the form of symbolic acts (SJmi jLu.:
D^n) and sayings (SJjs Sjl*.: Dnsi). 7

Co-ordination of the Quranic revelation with that process of Gemeinde-
bildung was the achievement of haggadic exegesis, in which the essen-
tially anonymous references of the text of revelation were carefully related
to the originally independent figure of the Arabian prophet. The haggadic
literary devices were many and varied. 8 The extent to which the haggadists
were concerned primarily to elucidate a fixed scriptural text has perhaps
been exaggerated. 9 To describe at least part of Ibn Ishaq's activity, for
example, as exegetical (tafsir) is convenient but, if the technical term is
construed in its traditional sense (explication de texte), possibly misleading.

1 Margoliouth, 'Omar's instructions', esp. 309-10.

2 Margoliouth, op. cit. 313. 3 See below, pp. 74-6.

4 Cf. the author's appropriate observations, 326; and see below, IV pp. 158-9.

5 See below, IV pp. 199-200.

6 Suyuti, Itqdn iv, 174; cf. above, I pp. 51-2; Goldziher, Studien ii, 20; Andrae,
Person, 179-80; Schacht, Origins, 149: the tradition was known to Ibn Qutayba but not
to Shafi'I.

7 Cf. Strack, Introduction, 9, 17; Gerhardsson, Memory, 82, 120-1.

8 See below, IV pp. 122-48.

9 Cf. Becker, 'Grundsatzliches', 520-7; Watt, 'Materials', 23-34; my observations
in BSOAS xxxi (1968) 148-9.

58

QURANIC STUDIES

In this context Sellheim's structural analysis of the Sir a deserves notice. 1
The author's discernment of a three-tiered composition (pp. 48-9) derives
from a separation of three kinds of material: the Hijazi environment
(Grundschicht: pp. 73-8), prophetical legend (erste Schicht: pp. 53-73),
and justification of the 'Abbasid dawla (zweite Schicht: pp. 49-53). The
taxonomy depends upon material distinctions rather than formal (stylistic)
ones, and appears to be at least partly informed by relation to the agreed
data of Ibn Ishaq's own career. As such it is a valuable contribution
towards solving the familiar problems about motives and materials in the
earliest stages of Islamic historiography. 2 Now, the Muslim concept of
Heilsgeschichte depended, not unexpectedly, upon the didactic value of
exempla, and those constitute in turn a major portion of scripture. 3
Whether such reflect Muhammad's idea(s) of history is irrelevant. That
they represent the organizing principle of Ibn Ishaq's composition is
relevant : the figure of an Arabian prophet, warts and all, might be thought
to provoke the question not so much of historicity as of faithfulness to the
traditional (Judaeo-Christian) concept of prophethood. From the point of
view of form- criticism Sellheim's Grundschicht may be a misnomer, if by
'basic' he would propose a contrast between a nucleus of historical 'truth'
and, on the one hand, the embroidery of prophetical legend (erste Schicht)
and, on the other, the transparent motives of political patronage (zweite
Schicht). All three of the structural levels exhibit a single impulse, namely,
a concern to locate the origins of Islam in the Hijaz. The emergence of
an Arabian prophetical tradition, of which the earliest agent appears to
have been the Sira of Ibn Ishaq, may well have contributed to its author's
dispute about methodology with Malik b. Anas and his subsequent depar-
ture from Medina. Indeed, an important problem in the analysis of the
Sira, and one only alluded to by Sellheim, is Ibn Ishaq's treatment of
material preserved also as the canonical text of revelation. 4

From the Quranic data themselves emerge several characteristics em-
ployed by the exegetes to establish a relation between the utterance of God
and its appointed recipient. These concern modes of revelation and came
to figure significantly in the literary elaboration of a biography appropriate
to the prophet of Islam. 5 As was noted, exegetical discussions of Q. 42: 51
stressed the equivalence zvahy.ilham, from which it was inferred that one
mode of revelation consisted in divine 'inspiration'. That justification for

1 'Die Muhammad-Biographie', 33-91.

2 Cf. Rosenthal, 'Influence', 35-45.

3 E.g. the employment of 'ibra by Tabari, Annates I/78, cited by Rosenthal, op. cit.
38-9; cf. Abbott, SALP i, 6-9.

4 Sellheim, op. cit. 47, 91; see below, IV pp. 127-9.

5 See above, I pp. 34-8.

EMBLEMS OF PROPHETHOOD

59

this inference had to be based on extra-canonical usage is clear from the
single (problematic) occurrence in scripture of alhama U I yiij Uj^>=s Lg^jti
(Q. 91: 8), itself the subject of dispute between two interpretations: in-
formed/communicated by God, or created/implanted in the soul by God.
In the lexicon of scriptural exegesis, as contrasted with that of philosophy,
it was the former interpretation which prevailed, attested as early as Kalbl,

adQ. 12: 15 «u^jf Jljuj (J-J/^j- J—jf ijl «*-J I l~s»-_?l?> though only
as one alternative to the rendering of awha as irsal (dispatch). 1 In later usage
ilham is unambiguously inspiration, e.g. in Suyuti ^Juil^J I S.UUJI <U|I
djyi. 2 What appears very likely to have been the source of the semantic
juxtaposition of wahy and ilham is reflected in a passage from the seventh/
thirteenth-century theologian Ibn Qudama '1-Maqdisi \jiyu 61 j^-.jJu ^
p^i! L5" <*-ic jJ3\j Jy ^1 J*« y- -Oil jV (sic) \1\J yt^JI 61
^53j5 J Jjj^*.. 3 It seems not unreasonable to regard ilham in the

sense of inspiration a borrowing from the terminology of (profane)
rhetoric and the several attempts (not supported by Zamakhshari and the
Mu'tazila) to preserve a distinction between ilham and wahy (as dispatch)
a reaction to the source of that borrowing and ensuing confusion. 4 Other
terms appropriate to the modes of revelation may be thought to corrobor-
ate this argument.

Even the word tanzil, a scriptural convention for 'revelation' (e.g. Q. 69 :
40-3), could be employed to describe poetic inspiration (scil. of Hassan
b. Thabit). 5 Of significance in this respect is collocation of the verb
tanazzala (descend) with satanic agents (shayatin) in Q. 26: 210, 221-2,
exhibiting the specifically Quranic imagery in which shay tan was identified
as an agent of evil. 6 That this was not always so is clear from the report,
adduced by inter alios Suyuti, according to which delay in revelation to the

prophet Muhammad was the result of desertion by his shay tan: ^JcJA

Tabari's reports on the satanic agents operative in the utterances of false
prophets reflect of course Islamic doctrine on the modes of revelation.

1 Tafsir, MS Ayasofya 118, izg r ; but cf. the same author's interpretation of azvhd
in Q. 12: 102 as akhbara, above, I p. 34.

2 Itqdn i, 164.

3 Goldziher, Abhandhmgen i, 6 n. 5; see also Heinrichs, Arabische Dichtung, 35-6;
cf. BSOAS xxxiii (1970) 616.

4 Pace Heinrichs, loc. cit.

s See Jeffery, 'Scripture', T90; Blachere, Histoire ii, 333; Shahid, 'Contribution', 573.

6 Shahid, 'Contribution', 569-72, 577-8.

7 Itqdn i, 91 ; Zamakhshari, Kashshdf iv, 765-6 ad Q. 93: 3 identifying the woman as
Umm Jamil/Jumayl, wife of Abu Lahab; for the several traditions in T aDar ', cf. Birke-
land, The Lord guideth, 13-8; and Bukhari: 'Der Schaitan der Propheten ist der Engel
Gabriel', cited Wellhausen, Reste, 134 n. 2.

6o

QURANIC STUDIES

The essentially neutral content of shaytan (daemonic as opposed to dia-
bolical) in the passages from Suyuti and Bukhari could be interpreted as
evidence of a situation in which prophetic and poetic inspiration were,
if not identical, at least closely related. 1 That the tale reported of the poet
Umayya b. Abi 'l-Salt, in which his breast was opened and filled (with the
gift of inspiration), should figure among the infancy stories of the prophet
Muhammad might be thought to have similar significance. 2 Negative
statements in scripture traditionally associated with the Arabian prophet,
such as Q. 69: 41 Jji Lj, 69: 42, Jji Vj, and 37: 36

jtXti U^Jt Ij5^jU) bj, as well as the celebrated attack on the

'poets' (26: 221-7), exhibit exceptive constructions in which the content
of the message rather than the source or mode of inspiration is im-
pugned. 3 References to the prophet as visionary (2 Samuel 24: 11), as
seer(i Samuel 9: 9), as mad (Hosea 9: 7), anyway contain such tradi-
tional imagery as seriously to diminish the impact of whatever invective

they might bear, perfectly expressed in Q. 51 : 52 ^ ^jjJI ^J? L> dU-iS"
Cij^-i j I ^L. VI Jj-^j Instructive examples of parallel

phraseology for divine and satanic inspiration are generated by Quranic
application of the verb alqd (literally to cast, but often synonymous with

arsala, or with awhd> in the sense of dispatch), e.g. Q. 40: 15 ^ l jL
o^J (cf. 4: 171; and a similar construction with nafakha 21: 91), 20: 39
JL* dLU oJiJfj and 8: 12 i_^tjJI \jy£~ '^jJI J i^t"- 4 The
imagery was perpetuated in the exegetical tradition, e.g. qjJU! <j Jii^j j\
<Olp «uHLj Jj^il (Jl i}yJ ^ji»JI describing the activity of Gabriel. 5 In
Q. 22 : 52, however, that same activity is attributed to a satanic agent ^1
<j\kJi}\ Jib L <U!l £-~us . . . «ic^ul J jlk~£JI ( cr - 22: 53. and 20: 87
with reference to al-sdmiri). 6 The centrality of Q. 22: 52 in discussions of
the Islamic theory of abrogation has been demonstrated by Burton, and it
seems clear that pertinent interpretation of satanic alqd required that it be
understood as counterpoint to ansa (to make forget), e.g. Q. 6: 68, 12:
42, 18: 63, 58: 19. 7 The significance in those discussions of satanic, as
opposed to divine, agency consists in the light it throws upon the semantic

1 Goldziher, Abhandlungen i, 107-8; cf. Jeffery, 'Scripture', 191 n. 10.

2 Goldziher, Abhandlungen i, 3, 213; Horovitz, 'Himmelfahrt', 171-3; see below, pp.
66-7.

3 Pace Shahid, 'Contribution', 568-73.

* Cf. Goldziher, Abhandlungen i, 5, 7 on nafahkalnafatha.

5 Suyuti, Itqan i, 125.

6 Cf. Geiger, Was hat Mohammed, 163-4; Horovitz, Untersuchungen, 114-15; and the
literature cited Paret, Der Koran, 334~7. esp- Yahuda, 'The Golden Calf and the Samiri'.

7 'Cranes', 253-4, 265 ; see below, IV pp. 195-6.

EMBLEMS OF PROPHETHOOD

61

evolution of shay tan: from poetic muse to God's particular adversary, from
shay tan zsjinni to shay tan as iblis. 1

The parallel and easily confused sources of prophetical and poetic
inspiration noticed here have an approximate Biblical counterpart in con-
flicting reports on the motive of the Davidic census, related in 2 Samuel 24:
1-9 and 1 Chronicles 21: 1-7 (cf. 1 Samuel 16: 14 on Saul and the two
spirits of God). Quranic adaptation of the Judaeo-Christian Satan will not
have been a consequence merely of antonomasia, nor yet of an attempt to
separate prophet from poet (for both might be divinely inspired), but
rather, of a persuasion that all inspiration required an intermediary. It may
also be observed that Widengren's description of the prophet as recipient
of revelation concedes but nominal recognition of this very characteristic
element in the Muslim concept of scripture by stressing inordinately his
movement towards and confrontation with God. 2 That imagery, too,
belongs to the exegetical tradition, but seldom, save in allegorical and
sectarian interpretation, impinges upon scholarly understanding of the
modes of revelation. Whatever body of prophetical 'wisdom' might from
time to time have been regarded as supplementary to the contents of
scripture, it was with an organized corpus of recognizable logia that the
mainstream of Islamic theology was concerned, and not with a source of
concealed wisdom for the elect. The appearance, at several levels of popular
and sectarian theology, of elements drawn from the inexhaustible pool of
Oriental Gnostic concepts is undeniable, but did not much influence the
stability of orthodox Muslim doctrine regarding the content and mode of
the Quranic revelation. 3

The agency of mediation is symbolized, somewhat ingenuously, in two

scriptural passages: Q. 6: 112 IjJ* ^ <J53 Li** dHjS'j

JjiJI J^j a* i .J\ ( »4^ « .> and 72: 27-8 ^ ^ VI

While the first postulates for every prophet (nabt) a Satanic tempter hostile

to the divine mission with which he has been entrusted, the second pro-
vides for a guardian angel (rasad) to ensure that the mission is fulfilled :

(^-isLJtJI {j* Jbjkibfc; 3£j^JI & -Vj- 4 This anthropomorphic ex-
pression of the sources of divine communication found significant elabora-
tion in Muslim views on the part played by the angel Gabriel in the process
of revelation. In Juwayni's typology it is made quite clear that Gabriel was
the agent of transmission for both Qur'an and Sunna, the former literally

1 Cf. Goldziher, Abhandlungen i, 7, 106-17; id. 'Ginnen', 685-90; Geiger, Was hat
Mohammed, 98-100; Horovitz, Untersuckungen, 87, izo-i; Ahrens, 'Christliches im
Qoran', 176.

2 Apostle of God, 124-7, 207-8. 3 Widcugren, op. cit. 100 n. z.
* Zamakhshari, KasMidf iv, 633 ad Q. 72: 27-8.

62

QURANIC STUDIES

(bil-lafz) and the latter conceptually (bil-ma'na). 1 While later exegetes
appear to have agreed upon the Qur'an as referent for the accusative pro-
noun in Q. 97: i, Barth's proposal that it was Gabriel dispatched by God
in Laylat al-qadr is not only likely from the point of view of syntax, but
is supported by Kalbl's interpretation of a similar construction in Q. 12: 2

iiJ tSyy-* jj* lMjaJIj Jj^- UJy'f Ul Jjib L^c UT/ aUyl U

Ajj^l. 2 Explicit mention of Gabriel in Muslim scripture (Q. 2: 97-8,

66: 4) would seem hardly to bear the burden of exegesis produced to depict
his central role in the mechanics of revelation: Zamakhshari found it worth

while to explain that the accusative pronoun in <UJ| j3L <iJJj> J^. i}y ju'U

referred to the Qur'an. 3 But paucity of reference could be and was com-
pensated for by identifying Gabriel with the Spirit, e.g. ruh al-qudus
(Q. 2: 253, 5: no, 16: 102), al-ruh al-amin (26: 193), ruhana (19: 17),
al-ruh (40: 15 etc.), an equation difficult to reconcile with Q. 17: 85
and the elaborate story of Muhammad, Gabriel, and the rabbis of
Yathrib. 4 In the light of both Biblical and Rabbinic allusions to
Gabriel, the Muslim allegation that he was regarded by the Jews as
an enemy poses something of a problem. Most of the reasons usually
adduced to support this contention are set out in Zamakhshari: (a)
that Gabriel had revealed to the Jewish prophet (sic, cf. Jeremiah 27)
God's intention to destroy the Temple through the agency of Nebuchad-
nezzar; (6) that God had commanded Gabriel to establish prophethood
among the Jews but he had taken it elsewhere (scil. to the Arabs) ; (c) that
he had revealed the secrets of the Jews to Muhammad. These details,
together with an account of an altercation between 'Umar and the rabbis
of Yathrib about the relative merits of Gabriel and Michael, were tenuously
attached to the Quranic phrase (2: 97): Say, who is an enemy of Gabriel?
Itself polemical in tone, the phrase might be thought to reflect the several
versions of a test of 'true prophethood' imposed by the Jews of Yathrib/
Medina upon Muhammad or by Quraysh with their assistance, in each
instance thwarted by the intervention of Gabriel on behalf of the Arabian
prophet. s This interpretation would point to the third of the three reasons
for hostility between Gabriel and the Jews, namely, that he had revealed
their secrets to Muhammad. The nature of the polemic must, I think, be
understood as exclusively Judaeo-Muslim, and the role of Quraysh seen as
a literary embellishment designed to show that opposition to Muhammad
was (also) Arabian. Gabriel's position in Muslim prophetology is, after all,

1 Cited above, I pp. 51-2.

* See above, I pp. 34-5; Barth, 'Studien', 119; Kalbi, Tafsir, MS Ayasofya 118, i28 r .

3 Kashshdf i, 169 ad Q. 2: 97.

4 See below, IV pp. 122-6.

5 Kashshdf, loc. cit.; cf. Suyuti, Itqan i, 97.

EMBLEMS OF PROPHETHOOD

63

not qualitatively different from that predicated of him in Daniel 8:15,9:21,
possibly 10: 9 if., and in Rabbinic literature. 1

In Muslim, as in Rabbinic, tradition one of Gabriel's primary functions
is that of pedagogue: as he had been guide and mentor to Joseph (Targum
Pseudo- Jonathan ad Genesis 37: 15; Talmud Babl. Sotah 36b) and to
Moses (Exodus Rabba I 67b), so too for Muhammad he performed the rites
of initiation into prophethood, instructed him concerning ablutions and
the times of prayer, guided him during his ascension to heaven, and
arranged for him the content of revelation during meetings in Ramadan. 2
For the Arabian prophet that instruction was of particular significance
since, according to the traditional interpretation of ummi in Q. 7: 157-8 he
was illiterate: ]ju i_»x5Cj V (^jJI. The manner in which this dogma

influenced discussion of the modes of revelation has been noticed. 3 The
consequent postulation of an equivalence ummi:' am ha-ares may reflect
a misunderstanding of Q. 2 : 78/ On the other hand, a similar dogmatic
impulse in Patristic literature, according to which Jesus and the apostles
were described as anthropoi agrammatoi, was adduced by Wensinck. 5
The basis in Christian scripture for that view, e.g. John 7: 15, Acts 4: 13,
might be thought to exhibit a specifically anti-Rabbinic (Pharisaic) ten-
dency. An analogous orientation in Muslim tradition is illustrated by
several elements in the Muhammadan evangelium. 6

The same Quranic passage (7: 157) provided a point of departure for the
allegation that the Arabian prophet had been prognosticated in Hebrew

and Christian scripture: ^aJss. Lj^C «*jj-^sj ^1 J^JI

J~*jVlj SljjJI J. Despite the further charge that sectaries of both

religions had falsified and concealed (tahrif, kitman) those parts of their
scripture which predicted the coming of Muhammad, the search for
proof-texts (testimonia) was notably successful. 7 Since the technique
itself had clearly been developed and refined in the crucible of Judaeo-
Christian polemic (cf. Luke 24: 27), it is of some interest that its earliest
attestation in Muslim literature should be an interpretation of Ahmad/
Muhammad as the Paraclete of John 15: 23-16: i. 8 On the other hand,
the classical loci probantes from Hebrew scripture were (curiously, in

1 Cf. Geiger, Was hat Mohammed, 12-15, 200; Katsch, Judaism, 85-91; Horovitz,
Vntersuchungen, 46, 107; Pedersen, EI, s.v. gjabra'il.

2 References in Wensinck, Handbook, 59; see above, I p. 37.

5 See above, I p. 36: munajjam opposed to jumla.

4 See above, pp. 53-4; cf. GdQ i, 14. 5 'Propheten', 191-2.

6 Cf. Koch, Growth, 88; Gerhardsson, Memory, 12-3; and see below, pp. 70-1.

7 See below, IV pp. 189-90.

8 Ibn Hisham, Sira i, 232—3 ; cf. GdQ i, 9-10 n. 1 ; and bibliographical references GAS
i, 289.

6 4

QURANIC STUDIES

the light of Christian figural interpretation of the imagery in Isaiah,
Jeremiah, etc.) not the prophetical books, but, rather, passages from
the Torah, e.g. Genesis 17: 20, Deuteronomy 18: 15, 33 : 2, all objects of
later and detailed refutation by Maimonides. 1 But in the later efflorescence
of that polemic references to Isaiah especially, but also to the other pro-
phets and of course the Psalter, were abundant. 2 The passages adduced
consisted of such as were traditionally recognized (by Jewish and Christian
exegetes) to contain obviously messianic symbolism, and their use by
Muslim polemicists displays familiarity with both substance and tech-
niques of Biblical exegesis. One example was the attention devoted to
numerical value of the letters in the name(s) of the Arabian prophet,
AHMaD (Q. 61 : 6) and MuHaMMaD (3: 144, 33: 40, 47: 2, 48: 29), the
results of which calculations were seen to correspond to (numerical)
equivalents in selected Biblical phraseology, e.g. Genesis 17: 20. Mention
of this device (hisab al-jummat) is attested in the earliest Quranic exegesis
and related there to Jewish practice (scil. ppHBU and tnoo , 3i). 3

In Muslim scripture itself both Ahmad (Q. 6

and Muhammad (33: 40 p53Uj

jj-jjJI pJU.j -UII Jj~y l jS3j) occur in contexts exhibiting distinctly

messianic imagery. The locution 'seal of the prophets', traditionally
interpreted as reference to the last link in a chain of prophetical election

(filJVl j»-T), can, if somewhat arbitrarily, be related to occurrence of
finite forms of the verb khatama in the sense 'to place a seal upon'
(Q. 2: 7, 6: 46, 36: 65, 42: 24, 45: 23). As such it was synonymous with
Quranic faba'a (e.g. 9: 93), and that equivalence (pjU-r^Ua) was in-
corporated into the exegetical tradition. 4 The eschatological significance of
Q. 33 : 40 is, however, unmistakable, and the verse might be understood
to constitute an exception to the attested principle that a prophet be elected
from within his own community: thus, 'Muhammad is not the father of
anyone of you, but rather the messenger of God and seal of the prophets'.
That Q. 33 : 40 contains one of the four occurrences in scripture of the
name Muhammad suggests a particular polemic, in which not only the
credentials but also the identity of the Arabian prophet was in dispute. The
caique proposed by Hirschf eld, fiflW (Haggai 2:23, cf. Jeremiah 22: 24),
required that the Arabic cognate be interpreted, indeed, as signet (fab')

1 Iggeret Teman, 36-7; cf. Steinschneider, Polemische Literatur, 326-7.

2 See Steinschneider, dp. cit. 325-9, 389-92, and separate entries nos. 2, 14/66, 105 ;
Goldziher, 'Polemik', 372-9; Schreiner, 'Zur Geschichte', 595, 599-601, 613, 625-8,
642-7.

3 e.g. Kalbi, Tafsir ad Q. 3: 7, MS Ayasofya 118, 29 r ; Muqatil, Tqfsir intro. and ad
Q. 3: 7, MS H. Htisnii 17, 2 r , 35 v ; cf. Bacher, Terminologie i, 127, ii, 27-8, respectively.

4 e.g. Zamakhshari, Kaskshdf iii, 544-5 ad Q. 33 : 40.

EMBLEMS OF PROPHETHOOD

65

indicating divine election. 1 That put forward by Horovitz, a<f>payls
( 1 Corinthians 9 : 2, cf. Romans 4:11) equated khdtam with musaddiq (corro-
boratio), that is, verification of earlier prophets and scriptures, a frequent
Quranic usage (e.g. 2:101, 3 : 81). 2 The teleological interpretation suggested
by Jeffery, re'Ao? vopov (Romans 10: 4, cf. Daniel 9: 24) was in harmony
with traditional Muslim exegesis, as well as attested in Manichean litera-
ture. 3 Only the last proposal would seem to do justice to the eschatological
flavour of Q. 33 : 40, and of 61 : 6 where, incidentally, it is Jesus, not 'he
who shall come after me', who is designated musaddiq.

Both verses contain the kind of material from which the Islamic Pro-
phetenkultus was elaborated, and might be thought to refute the view that
the latter was diametrically opposed to the Quranic portrait of an Arabian
prophet. 4 The inherent weakness of that view is its dependence upon a
clear distinction between a later, sectarian (even extremist) development of
the concept detos dv8pwiros, and an original, factual, and sober account of
the pious man summoned by God. So tidy a dichotomy is supported
neither by the content of revelation nor the chronology of early Islamic
literature. Despite protests of the type T am only . . .'/'I am nothing but
. . .' (e.g. Q. 7: 188, 18: no; the type is formulaic, cf. 26: 115 for Noah,
and 19: 19 for Gabriel!), the biography of Muhammad formulated in
exegetical literature cannot be said either to distort or to contradict scrip-
tural data on the words and deeds of prophets in general. Indeed, from
the point of view of a literary analysis, it can be argued that the principal
difference between the text of scripture and the Muhammadan evangelium
lies merely in the canonical status of the former. Thematic and exemplary
treatment of prophethood in the Qur'an was reformulated in the evan-
gelium (sunna/sira) as the personal history of Muhammad. 5

As in the classical literature of Hebrew prophecy, accounts of the
prophetical call in Muslim scripture begin abruptly with one or another of
the formulae of commission, and dispense with description of whatever
preparation may have preceded the call. 6 I should hesitate, however, to
concede e silentio there was no such preparation for the reception of revela-
tion. 7 For the pre-classical period of Hebrew prophecy evidence of such
was occasionally transmitted, e.g. for Samuel (1 Samuel 1 : 20-8, 2: 18-21,
3: 1-4), presumably in the original documents underlying the Elijah/
Elisha cycles, but especially for Moses (Exodus 1 : 8-3 : 1 ff.). 8 The infancy

1 Researches, 23; cf. Goldziher, 'Bemerkungen zur neuhebraischen Poesie', 724-6,
for the poetic use of khatam: misrah (Isaiah 9: 6).

2 Untersuchungen, 53-4. 3 'Scripture', 266-7. 4 Andrae, Person, 292-3.
5 See Horovitz, 'Biblische Nachwirkungen', 184-9, on tne rather facile equations of

Jensen, 'Das Leben Muhammeds', 84-97. 6 See above, I pp. 12, 23-5.

7 Pace von Rad, Theology ii, 50-9. 8 Cf. Eissfeldt, Einleitung, 349-56.

4836C75 D

66

QURANIC STUDIES

stories of Samuel and Moses have at least two motifs in common: election/
dedication at birth, and a mode of life designed to induce responsiveness
to the word of God. The importance of the Mosaic exemplum, which also
dominated Jewish prophetology, in both Muslim scripture and exegesis
has been noticed. 1 That the two motifs should figure in the evangelium
infantiae of the Arabian prophet is not unexpected, though it is of course
quite possible that the immediate Vorbild was not Moses. 2 It may, on the
other hand, also not have been Jesus. The typology of such motifs, as set
out by Andrae, may be understood to represent or to be drawn from a pool
of narrative ingredients traditionally appropriate to the lives of holy men. 3
The manner in which these could be adapted to a particular ambient
emerges from Kister's study of the descriptive terminology employed in
accounts of Muhammad's piety prior to his call. 4 Collation of Gospel
material with Islamic tradition, as undertaken by Goldziher, is obviously of
Tralue but could be misleading. 5 As a distinct literary type the Evangelion
was not restricted to the Christian canon, but represented the historiciza-
tion of logia traditions found not only in Biblical but also in Rabbinic and
Gnostic literature. 6 The Muslim term pertinent to that genre is mab'ath
(mission), in which historical development is symbolized in the thematic
polarity: promise-fulfilment. The ingredients of the Muhammedan
evangelium vary from one collection to another, but most had achieved
literary stabilization by the beginning of the third/ninth century.

Literary transmission did not necessarily entail a fixed order, and
fluctuation of three elements in particular has been remarked and analysed,
namely, the purification, the beatific vision, and the ascension/nocturnal
journey. 7 The Quranic evidence for each of those incidents in the life of
Muhammad is tenuous indeed, and such agreement as does exist in
exegetical literature on chapter and verse exhibits acceptance of several
arbitrary connections between scripture and the prophetical evangelium.
For example, the ritual opening of the breast, or purification, reflects un-
doubtedly a formula for the origin of poetic inspiration, and juxtaposition
of its Muhammadan version to Q. 94: 1-3 rests upon the semantic equiva-
lence shaqq batn : shark sadr. Now, Schrieke has argued persuasively that
such and similar rituals are almost invariably preliminary to an ascension
(confrontation with deity), indicating thus a syndrome whose internal logic
requires no scriptural support. 8 The exegetical (as opposed to historical)

1 See above, pp. 55-6, and I pp. 35-8; cf. Maimonides, Daldla ii, chs. 32-45, esp. 36.

2 Ibn Hisham, Sira i, 155-7, 235. 3 Person, 28-56. 4 'Tahannuth', 233-36.

5 Studien ii, 382-93; id. 'Neutestamentliche Elemente', 390-7.

6 Cf. Koch, Growth, 59-60.

7 See Bevan, 'Mohammed's ascension', 51-61 ; Schrieke, 'Himmelsreise Muhammeds',
1-30; Horovitz, 'Muhammeds Himmelfahrt', 150-83.

8 'Himmelsreise Muhammeds', 6-9; cf. also Widengren, Apostle of God, 80-95, '99-
216 on the ascension as a literary topos; Birkeland, The Lord guideth, 39-55, speaks,
however, of the theological distortion of 'an original experience of God in the Prophet's

EMBLEMS OF PROPHETHOOD 67
link between ritual purification and Q. 94: 1-3 Ux^jj ijjJ^dll ^ji^ ^jf
&j$> \y2o\ t^jJI iijjj idk*. is none the less a legitimate one, and was of
particular value in discussions of the dogma relating to prophetical perfection/
infallibility ('isTtta). 1 First articulated in Fiqh Akbar II (dated by Wensinck
towards the middle of the fourth/tenth century),the dogma exhibits elements
of both sectarian emphasis upon the qualities of the imamate and of
Rabbinic views on the kings and prophets of Israel. 2 Application to the
Quranic text enabled exegetes to identify the 'burden' {wizr) of Q. 94: 2
with apparent mention of earlier transgression (dhanb) and error (daldt) in the
life of the Arabian prophet (e.g. Q. 40: 55, 48: 2, 93: 7), necessitating in
turn postulation of the earliest possible date in his life for the act of puri-
fication. In the evangeltum itself the dogma found elaboration in the story
of the attempt by Quraysh to seduce Muhammad with offers of power and
wealth, to which naturally he did not succumb. 3 Similarly, the second ele-
ment in the syndrome of prophetical initiation, the beatific vision, may
exist independent of scriptural support, which, however, could be and
often was adduced in discussions not so much of Muhammad's prophetical
experience as of whether and when the faithful might be expected to see
God. 4 Those verses which were considered relevant to Muhammad's
vision, e.g. Q. 53: 11-18, 81: 19-25, 48: 27, were the object of extensive
and contradictory exegesis, resolved, save in the litteratim theses of the
mystics, by resort to compromise in the form of a spiritual vision (J 3j Jj
i_JlaJI cf. aVn rvxi).' The vision was intimately related to the third ele-
ment in the syndrome of initiation : the ascension. That ascension (mi'rdj)
and noctural journey (isra) exhibit fissiparous production from a single
tradition seems clear: their not quite consistently separate treatment in
exegetical literature betokens a concern for chronological development in
the evangeltum, in which the ascension was combined with the already
preposited ritual of purification. 6

The Quranic verse to which that exegetical tradition was invariably

attached is 17: 1 Jse^JI ,JI j»l J^c*«JI tAJ oJuao gjj tS-Xll jlt^
l^alVt. The anonymity of this reference was conceded only by Bevan. 7

own life' ; id. 'Legend', 6-12, distinguished the shaqq batn version as an 'investigation*
motif, separate from the purification and prior to the vocation.

1 Cf. Andrae, Person, 134-9; Birkeland, 'Legend', 42-7, on the transition shaqq: shark.

2 Creed, 193: articles 8-9, commentary 217-18.

3 Ibn Hisham, Stra i, 295-7; cf. Matthew 4: 1-11, etc.; and see below, IV p. 122.

4 Wensinck, Creed, 63-6.

5 Andrae, Person, 68-85; Widengren, 'Oral tradition', 258-60; Goldziher, Richtungen,
105-6; a rationalistic interpretation worthy of Noldeke may be read, appropriately, in
DeGoeje, 'Die Berufung Mohammed's', 1-5.

6 Schrieke, 'Himmelsreise Muhammeds', 14; Horovitz, 'Muhammeds Himmelfahrt',
1 74~S ; Birkeland, 'Legend', 54-60. ' 'Mohammed's ascension', 53-4.

68

QURANIC STUDIES

The alternative, namely, that 'abd can only be Muhammad, implies sub-
mission to an interpretation of all the Quranic data which, in my
opinion, has yet to be demonstrated. 1 Far from providing unambiguous

witness to the Arabian prophet, this particular scriptural image (tSj—l
}U «Jl*j) is employed, in but slightly varying forms, only to describe
Moses' departure from Egypt (Q. 20: 77, 26: 52, 44: 23; laylan may
reflect the imagery of Exodus 12: 20-34). 2 Moreover, the introductory

formula (_jjJI jUc^, is most probably of cultic origin and general applica-
tion. 3 Without specification of the terminal points in the journey
Lyaiil Ji^uvaJI (Jl J^-JI, probably a gloss, identification of 'abd

with Moses might be thought confirmed by the following verses (17:
2 ff.). 4 On the other hand and without exception, it is with Q. 17: 1 that
the isra\ and more often than not the mi'rdj, are linked in the exegetical
tradition. In the light of the clearly Mosaic formulation of Muslim
prophetology, that connection can hardly be described as arbitrary or
fortuitous, but may reflect as well a mixture of motifs. 5 Sudden transport
by the spirit of God from one place to another is a motif not uncommon in
the literature of prophetical expression (e.g. Elijah: 1 Kings 18: 12,
2 Kings 2:11, 16; Ezekiel 2: 12, 8: 3, 11 : 1, 43: 5; cf. 2 Corinthians 12:
2-4), and it may be that some such instance of divine intervention lay
behind the hymnic imagery of Q. 17: i. 6

That transposition of imagery, from what must have been in origin a
reference to the Mosaic exodus to an expression of ecstatic movement, can
have been effected only by means of the phrase 'from the sacred mosque to
the furthest mosque* which, I have noted, may be an exegetical gloss
designed to accommodate within the canonical text the ascension episode
of the Muhammadan evangelium. Allusion in Q. 17: 60 to a vision (Lj

(_X"UJJ 'izs VI iiJUjjt ^1 Uj jJI Ui*>.) and in 17: 93 to an ascension (Jij j\

1 Cf. GdQ i, 134-7, "» 85-8; Widengren, Apostle of God, 96-114; Schrieke, 'Himmels-
reise Muhammeds', 13 n. 6; Horovitz, 'Muhammeds Himmelfahrt', 160-1: unchar-
acteristically ingenuous, and ironic in the light of his further observation on another
identification (162), 'Dass auch die europaische Forschung sie bisher ohne Nachprufung
ubernommen hat, beweist nur, dass sie sich keineswegs uberall von dem Banne der is-
lamischen Tradition befreit hat'.

2 A related locution is used twice of Lot's departure from Sodom, Q. n: 81, 15: 65;
see above, I p. 8. * See above, I pp. 17-18.

* All of Q. 17: 1 was judged an interpolation by GdQ i, 136-7, though Weil's proposal
of a forgery was rejected, GdQ ii, 85-8.

5 e.g. the pseudepigraphic Assumptio Mosis, cf. Eissfeldt, Einleitung, 770-1; Mou-
barac's linking, Abraham, 59-60, of Q. 17: 1 with the Abraham traditions is in my
opinion unjustified, though it was of course a Meccan sanctuary tradition which facili-
tated interpretation of masjid haram and masjidaqsd as toponyms, cf. Abraham, 53-81. The
Biblical Abraham was not merely a seeker of God, but also a founder of sanctuaries,
e.g. Genesis 12: 7, 8, 13: 4, 18, 22: 9.

« Cf. GdQ i, 134 n. 7.

EMBLEMS OF PROPHETHOOD

69

t L-.J I <j) are hardly relevant to the content of 1 7 : 1 . Both are polemical and
the latter hypothetical, as is the ascension imagery in Q. 6: 35 and 15: 14.
In the exegetical tradition the sacred mosque was identified as the Ka'ba in
Mecca and the furthest mosque simply as Jerusalem. 1 In Jerusalem the
Rock (sakhra) might also be specified, but relation of the ascension as well
as the nocturnal journey to Q. 17: 1 would appear to support identification
of the furthest mosque with heaven. 2 A corresponding spiritualization of
the point of departure ( !) for both is discernible in statements ascribed
to 'Abdallah b. 'Abbas, according to which it was the spirit (ruh) of
Muhammad which made the journey from a point depicted not specifically
as the sacred mosque, but more generally as sacred enclave (haram). 3 But
a tendency in the opposite direction, namely, to fix the terminal points of
the journey at the Ka'ba (Mecca) and at the Aqsa mosque (Jerusalem)
attests to the political significance of Islamic sanctuaries and only inci-
dentally to the exegesis of Q. 17: i. 4 The link between revelation and the
evangelium was, however, not neglected. The celebrated tradition pre-
scribing three pilgrimages (-U.L~> VI JU^JI I^J V) appears in
Muqatil's discussion of Q. 17: 1, together with several stories about the
sanctity of Jerusalem, exhibiting conflation of masjid al-aqsd with sakltra
bayt al-maqdis. 5 One is tempted, if not quite constrained, to see in those
sanctuary traditions the origin of the israjmi'rdj story, imposed upon Q.
17:1 much in the way Surat al-Fil was made the peg for a similar sanctuary
tradition concerning Mecca. 6 If, indeed, the exegesis did not in both
instances precede the revelation, it would none the less appear to have
originated independently of the verses it purported to explain. That same
ambivalent relationship between scripture and interpretation holds for
much of the content of the Muhammadan evangelium.

Attached also to Q. 17: 1 is a characteristic example of prophetical
vaticinatio ex eventu (akhbar al-ghayb): Quraysh, appropriately astonished
by Muhammad's report of his nocturnal journey, challenged him to
describe a caravan of theirs at that moment returning to Mecca from Syria
(sic). This the prophet duly met, adding details of its leading camel
and predicting its arrival in Mecca next morning. 7 That the source of the

1 e.g. Zamakhshari, Kashshdf it, 647 adQ. 17: 1; cf. Horovitz, Untersuchungen, 140-1.

2 See Schrieke, 'Himmelsreise Muhammeds', 13-15; Horovitz, 'Muhammeds Himmel-
fahrt', 162-9: the Rabbinic/Christian concept of 'celestial Jerusalem'.

3 Zamakhshari, loc. cit.; Kalbi, Tqfsir ad Q. 17: 1, MS. Ayasofya 118, i56 r : possibly
to accommodate a report that the prophet had begun the night in the house of Umm
Hani' bint Abi Tllib.

4 See Kister, 'Three mosques', 173-96: the traditions adduce, inconsistently, both the
Aqsa mosque and the Dome of the Rock. Cf. Paret, Dor Koran, 295-6.

5 Tafsir, MS H. Hiisnii 17, I57 v -8 r .

6 Pace Birkeland, The Lord guideth, roo-l ; see above, I p. 42.

7 Ibn Hisham, Sira i, 402—3, and most commentaries ad loc, e.g. Zamakhshari, Kash-
shqfii, 647; see below, IV pp. 120-1.

7°

QURANIC STUDIES

anecdote is the evangelium, not the document of revelation, is clear from
its inclusion in treatises on the proofs of prophethood {dala'il al-nubumoa). 1
A product of haggadic exegesis, its function was primarily entertainment,
but the concluding formula 'and still they did not believe' signals its
incorporation into the mass of criteria assembled to distinguish the true
from the false prophet, i.e. fulfilment of the prediction (e.g. Deuteronomy
18: 22). The classic reference in scripture for the vaticinatio is of course
Q. 30: 1-4, the dominant interpretation of which provided not only evi-
dence of genuine prophethood but also comment on the course of Oriental
history.* The range of akhbdr al-ghayb includes another kind of utterance,
which might be designated figural (typological). An example was the
warning given by Muhammad's camel at Hudaybiya to halt and negotiate
at the perimeter of the Meccan haram, interpreted by the prophet as
manifestation of the force which had arrested Abraha's elephant. 3 That an
animal should have been appointedinstrument of God's will (cf. Numbers
22: 22-35) is in this particular context less significant than the fact that
Muhammad alone could understand and explain the camel's action to his
puzzled companions.

Situations in which the inscrutable wisdom of the prophet was demon-
strated are not uncommon. A familiar mise en scene is the confrontation
between Muhammad and the rabbis, in which the agency of Gabriel was
central. In one such episode Muhammad proved himself able, without the
aid of Gabriel, to confound the rabbis, namely, in the case of the couple
taken in adultery. The story is conventionally formulated in terms of a
prophetical test, based on two alternatives: if Muhammad elected to
punish the couple by flogging, public humiliation, and banishment (com-
prehended in the term tajbiya) he was clearly a king; if, on the other hand,
he sentenced them to death by stoning he must be a prophet. His first step
was to engage the Medinese rabbis in dispute, during which he could
display his superior knowledge of God's law. In one account it was Muham-
mad who got the rabbi 'Abdallah b. Suriya to admit that the stoning punish-
ment was attested in the Torah; in another it was the Jewish convert to
Islam, 'Abdallah b. Salam, who revealed the treachery of an unnamed rabbi
who had, during the dispute with Muhammad, held his hand over the
relevant passage in the Mosaic law. Muhammad thus triumphed, and the
couple was stoned at the gate of his mosque in Medina. 4 Traditions relating
this story to the ertswhile Quranic 'stoning verse' (ayat al-rajm) belong to
the principal loci probantes in discussions of the Islamic doctrine of

1 e.g. 'Abd al-Jabbar, Tathbit dald'il nubutmoat sayyidind Muhammad, MS Shehid
Ali Pasha 1575, 22'.

1 See below, IV pp. 144-5.

3 Ibn Hisham, Sira ii, 310; Waqidi, Kitdb al-maghdzi, 587; cf. Vermes, Scripture,
135-40-

+ Ibn Hisham, Sira i, 564-6; cf. Hirschfeld, 'Controversies', 109-16.

EMBLEMS OF PROPHETHOOD

7i

abrogation. 1 The Sitz im Leben of the story itself, however, is the Muham-
madan evangelium developed out of Judaeo-Muslim polemic. Elements
like the marital status (ihsari) of the accused couple and the rabbi's con-
cealment (kitmdn) of the pertinent portion of the Torah with his hand
exhibit a secondary stage of accretion, in which doctrinal significance pre-
ceded polemical value. The latter is reflected in its anti- Rabbinic flavour,
common to a wide range of anecdotes in the evangelium, and might be
thought to share a symbolic quality with the adulteress pericope in the
gospel of John (7: 53-8: 11). The intentions of the two stories are ad-
mittedly opposed: the action of Jesus was to supersede the Mosaic law,

that of Muhammad to revive it (4j J^pj AjI^Tj <U)I jJ L>J ^ Jjt Uti),
to rescue it from dereliction at the hands of faithless custodians. 2 None
the less, the anti-Pharisaic propaganda of the one is reflected in the anti-
Rabbinic propaganda of the other. That John 7: 53-8 : 1 1 was itself drawn,
perhaps later than formation of the Christian canon, from a pool of
narrative elements traditionally associated with opposition to established
authority has been proposed. 3 A theme common to all such material was
public demonstration of true prophethood, one which could hardly have
real significance outside the Judaic tradition. Halakhic elaboration of the
theme stressed two different but related aspects: the source of prophetical
authority and the extent of prophetical jurisdiction.* In its primitive and
unembellished haggadic form, however, the theme is essentially apolo-
getic, and reflects a widespread and popular literary type.

Importance of the Mosaic law is exemplified in yet another component
of the Muhammadan evangelium : the story of the prophet's abstention from
food sacrificed to idols. 5 Kister demonstrated that conflicting versions of
the tradition exhibit the entire range of feasible positions with regard to
the onset of Muhammad's perfection/infallibility (*ipna), and in particular
whether it began before or at the moment of his prophetical calling. With
the gradual crystallizing of Islamic orthodoxy the quality of 'isma was seen
to be conterminous with the life of the prophet, conforming thus to other
data in the evangelium. 6 But the substance of the argument about food was
susceptible of halakhic extension, as formulated in the complementary pre-
scriptions of Q. 6: 118 (eat of that over which God's name has been
uttered) and 6: 121 (eat not of that over which God's name has not been
uttered). The sacrificial ordinances were thus neither dissolved as in the

1 Cf. references in Wensinck, Handbook, 221-2; Goldziher, 'Usages juifs', 79; Hirsch-
feld, Researches, 137; GdQ i, 248-52; Schacht, Origins, 53 n. 4, 73-4, 191 n. 5.

2 Ibn Hisham, Sira i, 566.

3 See Metzger, New Testament, 223-4 and references 273 ; but cf. Derrett, Law, 1 56-88 ;
Torrey, Foundation, 149-50.

4 See below, IV pp. 192-6.

5 Kister, 'Bag of meat", 267-75.

6 Cf. above, pp. 65-7; Birkeland, The Lord Guideth, 28-32.

[

72

QURANIC STUDIES

Synoptic tradition (Matthew 15: 10-20, Mark 7: 14-23), nor figuratively-
transposed as in Pauline doctrine (1 Corinthians 8), but rather, epitomized
as in the Apostolic promulgations to proselytes (Acts 15: 20, 28-9, 21:
25). 1 In the Muhammadan anecdote not adjustment of but adherence to
the Mosaic law was stressed (e.g. Leviticus 17: 7), in particular by the
Arabian prophet, whose exemplary figure was being delineated for edi-
fication of the community. Now, discussions of their precise relation to the
Mosaic law were characteristic of sectarian literature emanating from
communities in the Judaeo-Christian environment, and inclusion of this
anecdote in the Muhammadan evangelium might be thought to reflect a
similar concern. Specific mention of idols in connection with dietary laws,

e. g . Q.22: 30 jtfj^i cx* o^-P* !*=~ u r 5 ^ (p. L ^ r^ 1 r 50 ^ J

can of course be, and often was, construed as allusion to the practices of
pagan Arabia. But here, as elsewhere in the historicization of prophetical
logia, even persuasive elements of local colour must be judged against the
possibility of assimilation from a literary source.

Like the Muslim canon, the Muhammadan evangelium applied directly
and graphically figures from Biblical imagery. In the story of the first public
recitations from revelation it is told how Quraysh, sceptical and stubborn,
taunt Muhammad with their refusal to listen to and understand his

proclamations. Thereupon was revealed the verse oT^aJI oly lilj

IjjX^ UUa. S^VU dy-*j± V ^jJI dLuj (Q. 17: 4s). 2 The setting is

contrived and the motivation transparent, but the required exegetical peg
was provided : the obdurate audience rendered victim of its own utterance.
Veils were placed over their hearts and deafness in their ears, and a screen
erected between themselves and the prophet. The Quranic imagery (Q. 17:
45-6, 18: 57, 6: 25, and in the mouths of the scoffers 41 : 5) is developed
round the motif 'hardening of the heart', condensed in 5 : 13 ^>ji3 UU=-j
3L*.U. In the literature of prophetical expression the classic example is
Isaiah 6: 9-10, where, during the commissioning of the prophet, the plight
of Israel is so described. 3 The rather ingenuous point of departure for the
anecdote in the Muhammadan evangelium, namely the assertion by Quraysh
that their hearts were veiled and their ears deafened, is contained not only
in Q. 41: 5, but also 2: 88 ijli. bjii \ji\Sj. The latter verse, however
was traditionally interpreted as reference to the Jews (cf. Leviticus 26: 41
Vwn oaaV), the transfer of imagery to Quraysh being a frequently attested

1 Pace Geiger, Was hat Mohammed, 197; cf. Katsch, Judaism, 131-4; Gerhardsson,
Memory, 314-18.

* Ibn Hisham, Stra i, 314-17.
3 Cf. von Rad, Theology ii, 151-5.

EMBLEMS OF PROPHETHOOD

73

device of the exegetical tradition. 1 The Biblical motif is also Mosaic, e.g.
Exodus 4:21 "D^TiN pins 'lifl (referring to Pharaoh), and Pauline exegesis
of Exodus 34: 33-5 linked the veil (mo») of Moses with the obduracy of
Israel (2 Corinthians 3 : 12-18 KaAt>/i/xa). 2 The dual function of the veil,
symbolic of both the hardened heart and the protected messenger of God,
found expression also in Muslim scriptural interpretation. In his com-
mentary to Q. 17: 45 Tabari explained hijab as kindn (veil, derived from
akinna in the following verse, cf. Q. 6: 25, 18: 57, 41: 5), or as satir
(screen) and glossed the entire locution 'screened from the people, who
could thus not see him' (jujjj }\j ^UJI tf. This additional inter-

pretative element was elaborated in later exegesis, where the hijab was seen
as a shield to protect the prophet from attempts on his life. 4 Interpretation
of hijab as satir (screen) might, conversely, be applied metaphorically, as in
Saadya's Arabic rendering of Psalm 88: 15 (inn *piD TDOn) as Sinn sV)
"|n»m. s In Q. 17: 45 the Mosaic Vorbild is unmistakable, and the
functional equivalence of hijabimasveh (Exodus 34: 33-5) merely another
instance of resort to traditional imagery in the elaboration of Muslim
prophetology.

The ambivalent relationship between scriptural data on the qualities of
prophethood and the material of the Muhammadan evangelium is especially
conspicuous in Muslim discussions of prophetical thaumaturgy. The latter
consists exclusively of not very convincingly adapted stereotypes of
miracles traditionally associated with men of God, and catalogued in the
dald'il al-nubuwwa literature. 6 The arbitrary assignment of Quranic chapter
and verse to those components of the evangelium has been noted. Despite
elaboration of a systematic test by means of which the acceptability of a
miracle could be determined, the theological relevance of that material
remained minimal. 7 Neither eliminated nor replaced, it was instead super-
seded by refinement of a dogma in which the document of revelation could
play a rather more central part, and by which the role of the Arabian pro-
phet could be assessed in terms of historical perspective. Adduced by
Jahiz, that dogma found rudimentary expression in the typology set out by

1 Cf. Geiger, Was hat Mohammed, 12; Goldziher, Richtungen, 175-6.
* Horovitz, Untermchungen, 30; Gerhardsson, Memory, 285-6; Ahrens, 'Christliches
im Qoran', 170.

3 Tafsir xv, 66.

4 e.g. Ibn Kathir, Tafsir iv, 314; Suyuti, in Tafsir al-Jala.la.yn, 386, both ad Q. 17: 47;
the story, without scriptural reference, is retailed in Ibn Hisham, Sira i, 355-6, where the
agent of the assassination attempt is identified as the wife of Abu Lahab ; cf. the commen-
taries to Q. 93 : 3, and above, p. 59.

5 Galliner, Saadja, xxiv, 8, 47, n. 7.

6 e.g. Abu Nu'aym and Bayhaqi, cited Andrae, Person, 57-91.

7 Cf. Andrae, Person, 101-3, citing Iji, Kitdb al-Mawdqif.

74

QURANIC STUDIES

Ibn Qutayba: the miracles (signs) of Moses were characteristic of an age
of sorcery (zaman al-sihr), those of Jesus of an age of medicine (zaman
al-tibb), that (sic) of Muhammad of an age of eloquence (zaman al-baydn). 1
Emphasis upon the single miracle of Muhammad, the book revealed to him
by God, might be thought to contradict the data of the Muhammadan
evatigelium, while attesting simultaneously to the general validity of
miracles as proof of prophethood. Application of that criterion was con-
sistent, even when tradition provided no explicit record of a miracle, as in
the case of Shu'ayb, e.g. Zamakhshari's unequivocal assertion adQ.y: 85
with reference to the locution ^5CTj ^ jc-j p5oe.U Ji- 2 Derivation of the

procedure from a Christian Vorlage is hardly necessary, despite the nature
of the material upon which the Muhammadan evatigelium drew. The judge-
ment of Horovitz, that the Christian origin of the infancy and other stories
was so pronounced as to preclude their employment by the Muslim adver-
saries of John of Damascus, is somewhat ingenuous. One could as easily
argue that the content of the Qur'an, which consists almost exclusively of
elements adapted from the Judaeo-Christian tradition, must have disabled
its sectaries in controversy with Jews and Christians. 3 The role of the
miracle as prophetical credential was of such currency in the formulation
of monotheistic religion as to make derivation from a single source a futile
exercise indeed. 4 Rabbinic efforts to circumscribe the admissibility of such
must be weighed against explicit pronouncements on the subject expressed
in terms of popular faith, for example, by Maimonides. 5

The thoroughly traditional character of Muslim polemic may be judged
from the protests of mortality and disclaimers of miracles in the text of
scripture itself. 6 What became, despite those assertions, the specifically
Muhammadan miracle is alluded to in passages containing a demand not
merely for credentials, but for written confirmation of the word of God,

e.g. Q. 4: 153 t LJI Ulxf j^JU J^J of J*f That

particular request is attributed to those already in possession of a scriptural
revelation (ahl al-kitab). Similar instances may logically be ascribed to
the same quarter, even when not explicitly stated, e.g. Q. 17: 93, 21 : 5, 43 :

1 Jahiz, Kitab Hujaj al-nubuzvwa, 145-6, but in a context in which the objective his-
toricity of prophetical biographies is critically examined; Ibn Qutayba, To? ml, 10.

2 Kashshdfii, 127: no prophet without a miracle; conversely, there could be no miracle
without a prophet, cf. 'Abd al-Jabbar, TanzOi al- Qur'an, 480 ad Sura 105.

3 'Zur Muhammadlegende', 41-9; anticipated by Schreiner, 'Zur Geschichte', 593-5;
apparently shared by Becker, 'Christliche Polemik', 437, and Wensinck, 'Propheten',
192-8. Conclusions drawn from the writings ascribed to John of Damascus are anyway
questionable unless restricted, as they seldom are, to an assessment of polemical tactics,
see my observations BSOAS xxxiii (1970) 391-3.

4 Cf. JefFery, 'Scripture', 132-3; Khoury, Les Theologiens byzantins, 89.

5 Iggeret Teman, 56; cf. Gerhardsson, Memory, 178, 213.
* See above, p. 65 ; and I, pp. 6-7.

EMBLEMS OF PROPHETHOOD

75

31, 47: 20, 74: 52. Collocation in scripture of 'prophethood and the book*
(v_>k£!]|j Zj~}\ Q. 29: 27, 57: 26) or 'the book, judgement/wisdom, and
prophethood' (ij~JI j |^*Jlj 3 : 79> 6:89, 45 : l( >) refers exclusively

to the Jews. 1 For verses of outspokenly polemical content, like Q. 35 : 40
t^Lo ljh5" j^aUJI >l (cf. 37: 157, 46:4), I am inclined to in-

terpret kitab as decree/authority (sultan, cf. 7: 71, 37: 156), rather than
book (scripture), a conjecture supported by Zamakhshari as well as by
Quranic usage. 2 Kitab as scripture is seldom differentiated in the Qur'an,
and exactly which scripture is meant can be elicited only from context. 3
The inherent ambiguity sensed by exegetes for many passages is reflected
in Abu 'Ubayda's gloss of dhaKka 'l-kitab (that book) in Q. 2 : 2 as hddha
'l-qur'dn (this Qur'an). 4

Attempts at closer definition of kitab, such as nasib min al-kitab (a por-
tion of the book, Q. 3 : 23, 4: 44, 51, 7: 37), do not in fact eliminate the
ambiguity, though in the exegetkal tradition those passages were inter-
preted as allusions to the Torah. 5 On the other hand, tafsil al-kitab
(analysis/explication of the book) is once (Q. 10: 37) expressly predicated
of the Qur'an, while the term tafsil is elsewhere (6: 154, 7: 145, possibly
17: 12) a reference to the Mosaic revelation. The locution kitab allah (book
of God), occurring nine times in Muslim scripture (five of which may well
mean 'decree'), shared in the exegetical tradition a similar ambivalence.
The alternative kitab allah and qur'an in variant traditions of 'Umar's
instructions to Abu Musa have been noticed. 6 A number of revealing
anecdotes are related of the same caliph, e.g. that retailed by Ibn Hazm,
according to which 'Umar was one day approached by the rabbi Ka'b
carrying a book (sifr !), who said to him: Here is the Torah, read it ! 'Umar
replied: If you are certain that that is what God revealed to Moses, then
I will read it night and day ! The variant reply 'Then read it during the
night and day', ill suited both to the preceding imperative in Ka'b's
utterance and to the spirit of the story, reflects the dogmatic impulse
responsible for another anecdote, in which Muhammad forbade 'Umar to
read the Torah. 7 Just before the caliph's assassination the same rabbi in-
formed him of his imminent death, indicating that he had found it pre-
dicted in the book of God, the Torah (oljjdl J»j y- <U!l <JcS~ <j «Jut). 8

1 Cf. Horovitz, Vntersuchungen, 73 : a conjectured caique of Hebrew tripartite scripture.
* Kashshdf iii, 617 ad Q. 35 : 40, iv, 640^37: 157; cf. Augapfel, 'Das "kitab" im Quran',
393-

3 See Kiinstlinger, 'Kitab', 238-47; id. 'Gottes-Schriften', 72-84.

4 Wansbrough, 'Periphrastic exegesis', 250, 256.

5 Cf. Kiinstlinger, 'Mathanf, 596-8.

6 See above, p. 57.

7 Kitab al-Fisal wal-milal i, 217; Goldziher, 'Polemik', 345.

8 Tabari, Armales I/2722-3 ; cf. BSOAS xxxi (1968) 614-15, and below, IV pp. 189-90.

7 6

QURANIC STUDIES

In a different context altogether, Ibn Hajar was constrained to observe
that the locution kitab allah could refer to the Quranic codex (mushaf). 1
Collective designation of scripture(s) as kutub allah (books of God)

occurs in enumeration of the signs, as in Q. 2: 285 juSo^Lij <U)b (j-*T JS"

jJu,jj a-x5~j (cf. 4: 136, 66: 13, and the pair, 34: 44 and 68: 37, in which is

stressed the plight of those not granted a scriptural revelation). In exegetical
literature the total number of kutub allah was given variously: 70, 104, 125,
163, the lowest figure being in all likelihood a reflex of 4 Ezra 14: 45~7- 2
'Books' as prophetical credential is a notion widely attested in Judaeo-
Christian literature, for which the paradigm was undoubtedly the Mosaic
revelation. 3 In view of the central role played by the figure of Moses in both
the scripture and prophetology of Islam, his relegation to the rank of sor-
cerer in Ibn Qutayba's typology of miracles is striking. So too, the dichotomy
between the Arabian prophet of the evangelium and the recipient of God's
final revelation, itself a miracle. Implicit in the typology is not merely the
initial exchange of roles, but also precedence of the book over the prophet.
It seems to me unlikely that such a development could have taken place
outside the tradition of Rabbinic Judaism, in which Moses the leader of
his people was succeeded by Moses the bearer of divine revelation.
Translation of the word of God into a written record (scripture) was an
essential element in the Mosaic tradition (Exodus 34: 27), perpetuated in
the imagery of classical Hebrew prophecy (e.g. Isaiah 30: 8, Jeremiah
30:2).+

Quranic reference to the word of God may also, and not unexpectedly,
be allusion to scripture, e.g. Q. 4: 46 ^ djij*> Ij^U jjjJI ^
-u^lj^, (similarly 5: 13, 41), in which the action explicit in tahrif could
only apply to the written word. 5 Conceptually related to kalim (words)
in those verses is qawl (speech) in Q. 50: 29 JjJd J^SJI JjLj U, as well as
kalimat (words) in 6: 115 JjL, V (similarly 10: 64) and kalam

(speech) in 48: 15 JUII {$S~ Ulu of dj-^y. (cf- 2: 75). In three of its
four occurrences kalim requires to be understood as scripture (viz. Q. 4 :
46, 5: 13, 41; the exception in 35: 10). Similarly all four occurrences
of kalam, which appears only in construct with allah (viz. Q. 2: 75, 9:
6, 48: 15, 7: 144, the last with an appropriate pronominal suffix). Qawl
may be interpreted as scripture not only in Q. 50: 29, but also 23: 68

1 Apud Suyufl, Itqdn i, 221.

2 See Abbott, SALPi, 54; Widengren, Apostle of God, 13^-46; Eissfeldt, Einleitung,
773-

3 Von Rad, Theology ii, 40—5.

4 Cf. Jeffery, 'Scripture', 137.

5 See below, IV pp. 188-92.

EMBLEMS OF PROPHETHOOD 77

and 39: 18, though it more often renders 'word of God' in the sense of
an expression of divine will (decree, e.g. 16: 40, 27: 82, 37: 31). In
that dynamistic sense kalimat is also employed, e.g. Q. 2: 37 (Adam)
and 2: 124 (Abraham), but more common in that usage is the singular
kalima, save for amr (utterance) the standard Quranic locution for God's
decree, whether retributive (10: 33, 39: 19) or creative (7: 137, 37:
171). 1 Thus may be understood also those passages designated by
O'Shaughnessy the 'Christological verses' (i.e. Q. 3: 39, 45, 4: 171).* It is
only in the context of Quranic kalima/amr that one may speak of a reflex
of the logos doctrine in Muslim theology, and that in a Philonic rather than
Trinitarian sense. 3 A possible influence of the Christian doctrine might,
however, be seen in sectarian phraseology, for example, interpretation of
kalimat al-fasl in Q. 42 : 21 as 'All b. Abi Talib, or in the extreme formula-
tions of the mystics on the creative power of the prophet Muhammad. 4
For the Ash'arite doctrine on the eternity of the Qur'an it is kalam allah
(scripture) which was the subject of controversy, and which may be dis-
tinguished from kalima, rather as in Christian theology logos and rhema
refer respectively to Christ and scripture. 5 In Hebrew scripture the
locution 'word of God' (mrp *13"T) is characteristic primarily of the pro-
phetic oracle, and its status constructus may be judged comparable to that
of Quranic kalam allah. 6 The semantic range of the Hebrew formula
includes scripture (e.g. Jeremiah 30 : 2, 36 : 1-2), but also the episodic
expression of God's will (e.g. 1 Samuel 3:7, 1 Chronicles 22 : 8), which
corresponds to Quranic kalimajamr. The specifically creative command of
Psalm 33:9 lasn ms Kin vn *ifiX Kin 'O may be compared with Q. 2 : 1 17

jj5Ci ^5" <U Jyb UU \jX tjjiji or the retributive utterances of

Isaiah 9 : 7 and Amos 1 : 2 with waqa'a 'l-qawl (the word fell) in Q. 27 :
82, 85. For the metaphorical sefatai of Psalm 89: 35 Saadya employed
Arabic qawli (my speech), not only to avoid the anthropomorphism but
also as reflex of a Quranic term for command. 7

The extraordinary position occupied by 'scripture' in Muslim propheto-
logy requires to be examined in the light of two doctrines commonly
interpreted as unique to the theology of Islam, namely, that the Qur'an
is inimitable and the word of God uncreated. Discussion of both turned
upon the form and content of kalam allah, which might seem to have taken
on a dimension out of all proportion to its role as prophetical bona fides.
While that role was never neglected, it would be more realistic to suppose

1 But cf. Baljon, 'Amr of God', 7-18: equating amr with Biblical 'esah.

2 Koranic Concept, 10, 24 if.

3 Pace Bouman, Conflit, 15-16; cf. Speyer, Erzdhlungen, 24-5.

4 Goldziher, Ricktungen, 304; Andrae, Person, 333-57: haqiqa muhammadiyya.

5 O'Shaughnessy, Koranic Concept, 62.

6 Von Rad, Theology ii, 87-8. ' Galliner, Saadja, xxvi, 50 n. 18.

78

QURANIC STUDIES

that the qualities of inimitability and eternity were formulated in the
attempt to secure a position within the Muslim community for the docu-
ment of revelation. The fact of canonicity, here postulated as sum of a long
and uneven process of Gemeindebildung, meant acceptance of scripture not
merely as evidence for the divine commission of one man, but also and
more especially, recognition of its authority in the life of the community.
Establishment of a historical connection between revelation and its reci-
pient was, on the other hand, not simply a corollary of canonization. 1 In
the preceding pages it has been argued that the historical portrait of the
Arabian prophet conforms to a pattern composed partly of the Quranic
data on prophethood, in character emphatically Mosaic, and partly of
motifs drawn from a narrative tradition typically associated with men of
God. The centrality of that portrait in the description of its origins formu-
lated by the Muslim community lies in the role of the prophet as paradigm
(sunna). The extent to which^the specifically prophetical Sunna represented
a refashioning of customary law and community practice is not easy to
determine. 2 The tendency to subsume as much as possible of juridical
precedent under the heading sunnat al-nabi was clear at the end of the
second/eighth century. For a number of reasons, adduced at the end of the
last chapter, the introduction of scripture as supplementary source of doc-
trine cannot have been earlier. The appearance of technical literature on
the dogmas of its inimitability and eternity was even later. Now, from
chronological indications alone it might seem that the document of revela-
tion had achieved canonical status without being defined either as inimi-
table or as uncreated, but rather by virtue of its association with the
prophet of Islam. I am inclined to interpret that link as one designed to
support not merely the claim of Muhammad to prophethood (reflected in
Quranic emphasis upon the Mosaic revelation), but also the claim of
revelation to an effective role in the life of the community (already regu-
lated by the prophetical Sunna). Of the three qualities predicated of
Muslim scripture it was undoubtedly that of faithfully preserved pro-
phetical logia which accounted for its acceptance as a source of doctrine.
That the logia, once collected and canonized, might be granted enhanced
status as the inimitable and uncreated word of God, would not appear to
have been either logical or necessary. Both qualities, however, may be seen
as reflexes of Rabbinic attitudes towards the Mosaic revelation, possibly
adopted and modified in the course of Judaeo-Muslim polemic.

In the document of revelation itself that polemic is exhibited in those
verses containing a demand for 'scripture' as sign, e.g. kitab, suhuf, sura,
qur'dn. That such demands originated in a Jewish milieu is occasionally
explicit (e.g. Q. 4: 153) and elsewhere a logical inference. If Christian, the

1 See above, I pp. 47-52.

1 Cf. Schacht, Origins, 180-9: legal maxims in traditions.

EMBLEMS OF PROPHETHOOD

79

reference will none the less have been to Hebrew scripture, equally
canonical and perhaps the only meaningful instance of specifically scrip-
tural revelation for adherents of that faith. 1 Attempts in the exegetical
tradition to ascribe those demands to the pagan Arabs of Mecca represent,
in my opinion, elements of a Nachdichtung designed precisely to illustrate
the Hijazi origins of Islam. 2 For that particular argument the exegetes
were able to call into evidence an undisputed 'fact' of Arabian history in
the seventh century, namely, the rhetorical accomplishment of the Arabs.
Similarly I would submit that the concomitant challenge to produce an
identical or superior scripture (or portion thereof), expressed five times in
the Quranic text (the tahaddi verses: Q. 2: 23, 10: 38, 11: 13, 17: 88, 52:
33-4), can be explained only within a context of Jewish polemic. In three
of those verses (Q. 10: 38, 11 : 13, 52: 33-4) the challenge: Then produce
a sura/ten suras/a hadith ( I) like it, follows immediately upon an allegation
of forgery (yaqulun aftardhujtaqawtcalahu), a charge which might be
thought to come from those familiar with scriptural revelation. Further,

the principal of the tahaddi verses, Q. 17: 88 <io.*i».| ^i)

jJ&u jjj'tj V oTjill iJud Jl«j of jj* appears in a context introduced
by a locus classicus for the celebrated 'rabbinical' test of prophethood (17 : 85
And they ask you about the Spirit). Finally, a paraphrase of the challenge

may be found in Q. 28: 49 L^u <5Jjt>f AJUt Xs- ^ v^S <J»
where the quality of excellence is predicated alike of Qur'an and Torah.

In what precisely the inimitability of Muslim scripture (i'jaz al-qur'an)
consists was the subject of long, ardent, and finally unresolved dispute. 3
Considered schematically, the material of controversy included arguments
based upon: (a) divine prevention of the prophet's contemporaries and
posterity from producing a successful counterfoil (sarfa: a position tradi-
tionally associated with the Mu'tazili Nazzam); (b) the contents of the
document, seen to include information about matters past, present, and
future which could not possibly have come by natural means to the un-
lettered prophet (akhbdr al-ghayb: thus the dogmatic emphasis upon
Muhammad's illiteracy 4 ); (c) its composition (nazm/ta'tif), an area ex-
tended to include not merely linguistic form but artistic structure in the
broadest terms possible, a consequence in part of rejection of the sarfa
argument, seen by some to reduce the Qur'an to the status of any pheno-
menon contra naturam (khdriq lil-'ada).

1 Gerhardsson, Memory, 226.

* See above, pp. 58, 62, 72; I pp. 36-7; and below, IV pp. 122-6: the 'rabbinical' test
of prophethood.

3 A synoptic view in Suyuti, Itqdn, naw' 64: iv, 3-23; cf. Schreiner, 'Zur Geschichte',
663-75 '> von Grunebaum, EI, s.v. I'djaz.

4 See above, pp. 62-3 ; I p. 38.

8o

QURANIC STUDIES

From those three basic orientations elements were combined and varied
with considerable ingenuity producing, save in the close rhetorical
analysis of Jurjani and, to a lesser degree, of Baqillani, a final synthesis
characterized by subjective expression of the document's unimpeachable
wisdom and clarity. 1 The slightly unrealistic and, in the end, unsatisfactory
nature of the sarfa argument, whose very terms were self-defeating,
followed from its precluding the test which itself would demonstrate the
inimitability of the Qur'an. Thus it was that references to content, with
appropriate stress upon the vaticinatio, and to form may be seen to repre-
sent the dogma as distilled from the tradition. Assessment of the content
of scripture tended to be ethical rather than aesthetic and expressed as
wonder that one book could contain so much, indeed all that might be of
utility to men. That its incomparable composition was not self-evident
seems clear from the amount of literature produced to support the argu-
ment that it was. In this field attitudes ranged from a scrupulous desire to
avoid the stigma of preciosity (takalluf) attaching to the Quranic style to
the express intention of finding there the archetypes of all rhetorical
device. 2 Views on the applicability of the rhetorical sciences to the problem
of i'jaz include the affirmation of Suyuti, that such could be apprehended
only by means of those sciences (ma'ani, baydn, badi'), the denial of Ibn
Hazm, that the word of God (kalam allah) could in any manner be com-
pared to human utterance (kalam al-makhluqin), and the carefully qualified
statement of Baqillani, that while a variety of rhetorical embellishment was
exhibited in scripture, such must be irrelevant to the fact of its inimi-
tability.*

The significance of the i'jaz controversy may, I think, be sought else-
where than in its theological implications. In the course of those dis-
cussions it was, naturally, asked whether the Torah and Gospel(s), being
the word of God, did not share with the Qur'an the quality of inimitability.
The reply, not surprisingly, was no : (a) though like the Qur'an they did
contain reports of the unknown (akhhar al-ghayb), there was nothing of
the miraculous in their style (nazm) or their structure (ta'lif); (b) this
because God had not described them as such, nor had the challenge
(tahaddi) been referred to them, nor did the language(s) in which they were
written contain anything of eloquence (fasdha) ; (c) and finally, because no
such claim had been made for those books by their sectaries. 4 A further
point, attributed to Zarkashi and Zamakhshari, stressed the Quranic

1 See Weisweiler, 'Unnachahmlichkeit', 77-121; von Grunebaum, Tenth-Century
Document; Bouman, Conflit; Suyuti, Itqdn iv, 11-7.

2 e.g. the dispute about madhhab kalami, see below, IV pp. 232-3 ; cf. von Grunebaum,
Tenth-century Document, xv on the Venerable Bede.

3 Suyuti, Itqdn iv, 186; Ibn IJazm, Kitdb al-Fisal iii, 18-19; Baqillani, I'jaz, 107, 111-
12; von Grunebaum, op. cit. 49, 54-5; Bouman, Conflit, 71.

4 Baqillani, I'jaz, 31-2; Bouman, Conflit, 66.

EMBLEMS OF PROPHETHOOD

8z

arrangement (tartib) in suras according to divine plan, while such divisions
in the texts of other scriptures were the work of their compilers. 1

Of those arguments only one might be thought to contain some objective
basis outside the immediate terms of the controversy, namely, the con-
viction that eloquence was peculiar to the Arabic language. Introduction
of that element into what purported to be theological doctrine served a
dual purpose: Muslim prophetology was enriched by the concept of a
lingua sacra, and the elaboration of Islam characterized by its exclusively
Arabian origins. Exegesis of the tahaddi verses, in order to underline the
complete failure to meet the challenge (mu'drada: the absence of logic in
that reasoning was never admitted), presupposed the specifically Arabic
eloquence of all contestants. 2 When these failed, as indeed they must, how
could the claim to inimitability be denied? Rather more important than
the challenge which, paradoxically, must have been addressed to those
familiar with scripture, was the notion that the word of God could be
bound to a particular medium. For a number of reasons such a view was
hardly possible for Christians, though evidence of an emphatically con-
servative attitude to the Biblical text was not wanting even there. 3 On the
other hand, Rabbinic (and Targumic) designation of Biblical Hebrew as
the lingua sacra (leshon ha-qodesh) might seem to have provided the imme-
diate Vorbild for Muslim usage. 4 Now, there is admittedly no equivalent
in Rabbinic theology to the dogma of i'jaz al-qur'dn, and Jewish polemic
concerned with that doctrine appears to have taken as point of departure,
by calling into question, the Muslim premiss that the Quranic style was
demonstrably superior to that of profane compositions in Arabic. 5 It
might indeed be supposed that no other course was open to those whose
notion of lingua sacra was not bound to Arabic: aesthetic assessment of
the Quranic style was almost necessarily a preoccupation internal to the
Muslim community. Two aspects of the dogma which, however, could
have been devised to meet the needs of external polemic were the fact of
a sacred language other than Hebrew, and its appropriate bestowal upon a
people whose appreciation of rhetorical niceties was established. 6

A third aspect of the i'jaz controversy generated an independent series
of dogmatic propositions. In retailing Mu'tazili views on the inimitability
of the Qur'an Ash'ari cited the insistence of Nazzam on the sarfa argument
and added that both Hisham Fuwati and 'Abbad b. Sulayman refused to

1 Apud Suyuti, Itqdn i, 186-7; cf- the opposition munajjam : jumla in discussions of the
mode of revelation, above, I pp. 36-8.

2 e.g. as in Ibn Qutayba, Ta'zvil, 17; Bouman, Conflit, 66-73, seems here to have
missed the point of the argument.

3 See Loewe, 'Latin Vulgate', in CHB ii, 106-7.

4 Cf. Segal, Mishnaic Hebrew, 2-3.

5 Cf. the fragments in Steinschneider, Polemische Literatur, 103, 314 n. 23.

6 See below, III pp. 93—9.

82

QURANIC STUDIES

acknowledge the Qur'an either as witness to the existence of God or as
credential for the prophet. 1 While the sarfa argument could be inverted
to demonstrate the fact of Arab eloquence (in the absence of which divine
prevention of an imitation would have been unnecessary!), the rejection
of revelation as miracle provoked a different kind of rebuttal. The rejection
was based upon an assertion that the word of God, hence the Qur'an (sic),
could only be described in relation to the existence of God as contingent
('arad) : the Mu'tazili definition of God's unity (tawhid) precluded adducing
contingencies as witness either to His being or to His actions. Now, it is
generally assumed that the Mu'tazili position represented a reaction to an
earlier and more popularly held view that the word of God, hence the
Qur'an (sic), was neither contingent nor created but, rather, inseparable
from the unqualified existence of its creator. But it is difficult indeed to
discern from the expression of Mu'tazili views, most of which are preserved
only in the later works of their triumphant adversaries, whether they
exhibit opposition to other views already expressed, or merely the articula-
tion of conclusions derived from their own creed. Worthy of remark is
that the three Mu'tazili spokesmen cited by Ash'ari all lived in the first
half of the third/ninth century, and that the utterances ascribed to them
were made with specific reference to the inimitability of the Qur'an. It
might be thought that acceptance (or rejection except on the condition
of sarfa) of its inimitability somehow entailed definition of the Qur'an as
uncreated. As I have indicated, both qualities appear to have been formu-
lated with less concern for the description of prophetical bona fides than
for assertion of the document's canonical status within the community.
It may of course be argued that the evidence for a connection of the two
dogmas is too circumstantial to be of real value, and that, in any case, the
relation between the two cannot be described as causal. Ash'ari noted that
most members of the Mu'tazila accepted the inimitability of the Qur'an
(defined in terms of its composition!) without, apparently, admitting that
it was uncreated. 2 The converse could hardly have been postulated.

In Fiqh Akbar I there is, expectedly, no mention of the uncreated
Qur'an. In the Wasiyat AHHanifa, dated by Wensinck towards the middle
of the third/ninth century, the dogma is set out in Article IX, but without
mention of inimitability. 3 Those data accord with what seems to have been
the chronological framework of the controversy, from which emerged an
instructive terminus ad quern; c. 235/850. 4 The opinions of the dissenting
Mu'tazila cited by Ash'ari had been expressed at about that date. The
harmony between these details and the chronology proposed for the pro-

1 Maqdldt al-Isldnriyyin, 225-6; cf. Bouman, Conflit, 19.

1 Maqdldt, loc. cit.; an example was the Mu'tazili *Abd aL-Jabbar, cf. Bouman, 'Doc-
trine', 67-86.

3 See above, I p. 44; Wensinck, Creed, 127, 149-51, 187.

4 Cf. Watt, 'Early discussions', 29 n. 5, 33, 36.

EMBLEMS OF PROPHETHOOD

83

cess of canonization ought not to be overlooked. Now, that process might
be described as the collocation of several elements: a corpus of pro-
phetical logia, the figure of a prophet, a sacred language, and an un-
equivocally divine sanction for all three. Reconstruction of the manner in
which these elements were evolved and adjusted to produce a more or less
final and satisfactory synthesis can only be conjectural. It may be clear
from the preceding observations that I regard the acquisition within the
community of these elements as having taken place in the sequence set out
above. Their interaction might be envisaged as follows: attribution of
several, partially overlapping, collections of logia (exhibiting a distinctly
Mosaic imprint) to the image of a Biblical prophet (modified by the material
of the Muhammadan evangelium into an Arabian man of God) with a
traditional message of salvation (modified by the influence of Rabbinic
Judaism into the unmediated and finally immutable word of God).

Achievement of the final stage in the process may be equated with the
declaration that the Qur'an was uncreated. Identification of the Qur'an
document with the word of God (kaldm alldh) was derived from a traditional

metaphor, e.g. Q. 9: 6 ajl ^'j^S' U. A further identification of

the Qur'an document with the word of God as interpreted in the discussions
of divine attributes (sifdt) 1 appears to entail a transition from metaphorical
to veritable sense, which may be traced to the imagery of Q. 85 : 21-2
Jiji?w £-^J (j Xsy^, jl y Jj. The concept of a celestial archetype evoked

by this verse, which was crucial in the sifdt discussions, 2 belongs to an
ancient and well-attested tradition, in which of course the referent was the
word of God as injunction, law, even register, but not 'scripture' in the sense
of record of revelation. 3 From the notion of 'law' inherent in the use of
lawk (tablet) it is at least conceivable that qur'an in Q. 85 : 21-2 could be
functionally equated with the Mosaic law, and more particularly with the
Rabbinic concept of the pre-existent Torah as the immutable word of
God. 4 That the ultimate authority of God's decree might thus be fixed
provided a more or less satisfactory solution to the problem of scripture
as normative in the life of the community. But theological difficulties
provoked by the possible charge of dualism remained, and attempts to
solve or to evade these are evident in the works of masoretic exegesis, e.g.
Abu 'Ubayda's interpretation of Q. 85: 22 as ft luhin mahfuzun. 5 The

1 e.g. Ash'ari, Maqdldt, 582-607; cf. Goldziher, 'Fachr al-din al-Razi', 245-7.

2 Cf. Ash'ari, Maqdldt, 549 ff .

3 Cf. Horovitz, Untersuckungen, 65-6; Jeffery, 'Scripture', 51-3; but also Widengren,
Apostle of God, 115-61: for the equation Wisdom-Tablet-Book.

4 e.g. Talm. Babl., Pesahim 54a, Nedarim 39b, and the Midrashic references in Speyer,
Erzahhmgen 34 n. 2; see above, I pp. 35-6.

s Wansbrough, 'Periphrastic exegesis', 252 no. 29; cf. Zamakhshari, Kashshdf iv,
733 ad loc.

8 4

QURANIC STUDIES

reading 'preserved in an atmosphere' might be thought analogous to
Saadya's theory of avoir sheni, formulated to avoid the implications of a
hypostasized word of God. 1 In contrast with the masoretes, Muslim
theologians found satisfaction in distinguishing between the eternal word of
of God as concept (ma'na) and its reproduction (daliljtanzH), an argument
derived from a series of grammatical oppositions: qird'a-maqru', kitdba-
maktub, dhikr-madhkur. That formulation appears to have been the work
of Nazzam (d. 231/846) or Ibn Kullab (d. 240/854), incorporated finally
into the theories of Baqillani (d. 403/1013). Its refutation by the late
Mu'tazill 'Abd al-Jabbar (d. 416/1025) was unconvincing. 2

Such problems, with their solutions, remained peripheral. Textual
explication of Muslim scripture was able to accommodate variant, emenda-
tion, anomaly, ellipsis, pleonasm, in short all that might be described as
characteristic of any literary record, irrespective of convictions about its
source. The most interesting aspect-by far of the application of literary
criticism to the text of revelation was the extrapolation, from an allegedly
profane tradition, of aesthetic criteria formulated to describe the substance
and function of a sacred language. That was the product of masoretic
exegesis. 3 The Muslim masoretes were above all grammarians, whose
primary task was to relate the anomalies of a lingua sacra to the demands of
a normative description of language. With analysis of that process the
following chapter is concerned.

1 Altmann, 'Theory', 20-4.

2 Bouman, Conflit, 20-3, 37-8, 73-80; id. 'Doctrine', 74-84.

3 See below, IV pp. 202-27.

Ill

ORIGINS OF CLASSICAL ARABIC

Historical descriptions of the Arabic language are reminiscent of a
celebrated discussion among Renaissance humanists about the intrusion
of Italian into areas allocated by tradition to classical Latin. Against the
isolated conjecture that the literary (Latin) and vernacular (Italian)
languages were coeval but functionally separate was ranged a series of
arguments concerned to describe a developmental relation between sermo
urbanus and sermo plebeius: the latter was the product either of barbarians'
misuse of the former, or of local (indigenous) deformation; conversely,
sermo urbanus might be described as a conscious (social and aesthetic)
refinement of sermo plebeius. 1 The hypothesis of a functional dichotomy,
to be qualified by the observation that usage varied not with social position
but with the demands of a given linguistic situation, was both non-
evolutionary and not very forcefully asserted. The three evolutionary
concepts appeared, on the other hand, to make some sense of the historical
data, and have survived as points of departure for those theories of linguis-
tic genealogy which have not been entirely discredited.

Nowhere have such proved more durable than in studies of the position
occupied by Classical Arabic in the historical development of Semitic
languages. An inclination to see in Classical Arabic (CA) at least the
phonological and morphological constituents of a hypothetical Proto-
Semitic is both understandable and of some value for comparative and
diachronic analyses. To draw from the same data conclusions about the
origins and evolution of CA involves implicit acceptance of considerable
non-linguistic material often and erroneously supposed to be 'historical
fact'. I refer to such assumptions as that of the isolation of speakers/writers
of Arabic within the Arabian peninsula up to the seventh century, or that
of the existence of a ne varietur text of the Islamic revelation not later than
the middle of the same century. Both the origin and the utility of those
assumptions are patent. Evidence of abrasion (phonological/morphological)
and of interference (syntactic/lexical) could be ascribed to the use of CA
by foreigners or to widespread dislocation within the Arabic-speaking
community, both being consequences of the expansion of Islam during the
seventh century. Moreover, that CA could survive to serve as model
(orthographic/grammatical) for a literary language could be explained by

1 Vossler, Einfuhrung ins Vidgarlatein, 48-53.

86

QURANIC STUDIES

the fixed, sacred, and immutable text of the Qur'an, representing the
highest form of rhetorical achievement in Arabic. Provision of a geo-
graphical and chronological setting for the diachronic examination of
CA enabled observers of the purely linguistic phenomena to interpret these
as evidence of a more or less uninterrupted process of decline: movement
away from a point of linguistic (and ethnic!) purity towards a situation
characterized by fragmentation, dialect cleavage, unattainable ideals, and a
permanent tension between theory and practice. The historical assumptions
underlying such an interpretation of the data are, however, not merely
unverifiable, but also internally inconsistent and, in some respects,
demonstrably false.

Now, the marshalling of phonological evidence (consonantal range/
vocalic quality and position) as well as of that pertaining to morphology
and, in the strictest sense, to grammar (inflexion: case/mood) might be
thought unexceptionable, so long as it is recognized that such material
reflects the highly specialized and often idiosyncratic usage of rather
meagre literary records preserved in comparatively late recensions. 1
Syntactical evidence has been interpreted as exhibiting the effects of a
normative process, and thus as proof of some distance between CA and
Proto-Semitic, whose proximity, on the other hand, had been inferred
from the witness to phonology and morphology. 2 There is, of course, no
compelling reason why any of that linguistic material should yield con-
clusions of a chronological, as opposed to structural, nature. In the context
of evolutionary interpretation, the work of Vollers with reference to the
consonantal range of CA as exhibiting, in relation to Proto-Semitic, not
preservation, but rather, phonetic proliferation, and to the essentially
euphonic and rhetorical significance of the i'rab phenomena, represented
a reaction to traditional views, but required at the same time, in order to
fit the geographical/chronological environment alleged to have produced
the earliest Arabic literature, a very questionable reconstruction of the
history of the Quranic text. 3 The concept of CA as a kind of linguistic
(and literary) canon, apart from reflecting a well-attested and strongly
entrenched tradition of rhetorical criticism, may owe something of its
putative authority, at least for modern philological scholarship, to its
role as confirmation cum source of Proto-Semitic reconstructions. For
identification of actual specimens of CA, exponents of that scholarship
have for the most part been content to accept the received tradition as
found in works of rhetoric and exegesis, namely, that CA is exhibited
in the corpus of poetry described as pre-Islamic and in the document of

1 e.g. GVG i, 33-4, I20-I, 459-62. 554; Moscati, Comparative Grammar, 14, 52,
95-6, 134-5.

2 e.g. Bergstrasser, Einfuhrung, 134-5; Blau, 'Problems', 1.

3 'Arabisch und Semitisch', 165-217; id. Review: Noldeke, Zur Grammatik, 125-39;
id. Volksspracke, 55-175.

ORIGINS OF CLASSICAL ARABIC

87

Islamic revelation. The only deduction possible was that, within the
bounds imposed by available literary remains, CA stood at the beginning
and not at the end of or at a point in the course of a long and varied
linguistic evolution. 1 To interpret that evolution, which seems always to
to have been identified with description of the vicissitudes of CA in a post-
classical environment, two criteria in particular have emerged with an
apparently reasonable claim both to utility and to objectivity.

The first of these may be described as a cluster of interpretative prin-
ciples derived from the concept of koine, adduced to provide a terminus
ad quern for the historical evolution of CA. Koine has been employed to
designate: (1) the language of pre- Islamic poetry and Quran, (2) a kind
of bedouin lingua franca, and (3) the hypothetical source of modern
sedentary vernaculars. 2 Despite the inadvisability of accepting one term
for such varied phenomena, the several applications of koine reflect a
certain unity of impulse, namely, to describe difference in terms of diver-
gence from a single source, an intellectual principle commonly associated
with genetic linguistics. Though the implications of koine taxonomy for
the history of Arabic, particularly in respect of its second and third uses,
have been questioned and eventually modified, the definition of CA as
the language of poetry and of revelation has not been unseated. 3 The koine
principle was also applied to the scanty and problematic evidence of dialect
cleavage to produce a series of plausible if necessarily hypothetical polari-
ties: e.g. regional (Nejd-Hijaz), economic (nomadic-sedentary), social
(patrician-plebeian), ethnic ('arab-'ajam). One result of the experiment
was to demonstrate that, contrary to the assertions of Muslim philologists,
CA was not and had never been identifiable with any single Arabian
dialect. 4 By a somewhat eccentric application of logic, that negative
argument confirmed the identification of CA with the ineptly formulated
'poetic koine'.

The second criterion adduced to describe the fortunes of CA includes
several variations upon the theme of pseudo-correction. As a descriptive
principle pseudo-correction presupposes the existence of an acknowledged
standard of linguistic (or literary) excellence and witness, across a range
of individual speakers/writers, to incomplete mastery of it. 5 Thus, the
evidence of pseudo-correction cannot alone be employed to demonstrate

1 Which did not, of course, preclude versions of a hypothetical pre-history, cf. Rabin,
'Beginnings', 35-6.

2 Cf. Blachere, Histoire i, 79-82; Fuck, 'Arabiya, 7; Noldeke, 'Das klassische Arabisch
und die arabischen Dialekte', BSS, 13; Ferguson, 'Arabic koine', 616-30.

3 Cf. Cohen, 'Koine', 119-44; Blau, Emergence, 10-18; GVG i, 23; Spitaler, Review:
Fiick, 'AraKya, 145.

* e.g. Rabin, West-Arabian, 17-24; id. 'Beginnings', 32-4: the view that CA could
be traced ultimately to a dialect was, incidentally, not the exclusive property of Muslim
scholarship.

5 See Blau, Pseudo-corrections, 1 1-22.

88

QURANIC STUDIES

existence of a linguistic/literary standard, proof of which must already
have been established. For the history of CA it need hardly be added that
applications of this criterion will not be very helpful in determining just
when that particular form of the language attained its status as standard. 1
But the usefulness of the principle is not thereby exhausted. Absence of
pseudo-correct features may indicate either mastery of the standard or,
alternatively, that the standard is not distinguishable from the usus loquendi.
The two possibilities may, it seems to me, be regarded as mutually ex-
clusive, evident from the composition of works in faultless CA by authors
who could not conceivably have spoken that language. To insist that the
absence of pseudo-correct features in the Qur'an (accepting that such is
the case) reflects not mastery of the literary idiom but rather the 'true
vernacular of Mecca' involves an assumption based on non-linguistic
evidence, namely, that at the time the Qur'an was recorded in writing
the literary and spoken languages were, at least in Mecca, indistinguish-
able. 2 Necessarily conjectural, that assertion would appear also to pre-
clude an alternative possibility: that when the Qur'an was recorded in
writing the grammar of CA had been formulated and could be learned by
someone whose mother tongue it was not. Chronological problems are,
of course, not thereby solved, but neither are they prejudged by tacit
acceptance of the ' Uthmanic recension traditions. The criterion of pseudo-
correction, like that of the koine concept, has served iilogically to confirm
the position of CA in seventh-century Arabia by dependence upon it as a
historical fact. As instruments of synchronic analysis both criteria are of
indisputable value; for purposes of historical description they are at best
convenient fictions.

For the diachronic study of CA linguists have naturally, if somewhat
ingenuously, had recourse to the framework supplied by historians of the
early centuries of Islam. In their construction of that framework historians,
in turn, have relied largely upon a corpus of literature extant only in
recensions dating from the beginnings of the third/ninth century. Excep-
tions to that circumstance are few, indeed, the major one being, in the
consensus of Muslim and Orientalist scholarship, the text of the Islamic
revelation itself. General assent to 'le fait coranique' 3 links in an extraordi-
nary manner the most disparate and even contradictory interpretations of
early Islamic history, a secondary consequence of which has been affirma-
tion of the 'fact* by virtue of repeated assertion. In an essay which did not
depart significantly from the traditional view of the Hijazi origins of
Islam, the historian C. H. Becker proposed a distinction between the

1 Consider the dates of the material so carefully analysed by Blau in his Grammar of
Christian Arabic, 2-36; cf. BSOAS xxxi (1968) 610-13.

2 Cf. Blau, Pseudo-corrections, 57-8 esp. n. 15.

3 Blachere, Histoire ii, 230-41.

ORIGINS OF CLASSICAL ARABIC

89

processes of Arabicization and Islamization, in his opinion separated by a
period of from two to three centuries. 1 The former term was applied to
the several religiously neutral factors thought to have effected the move-
ment of the Arabs beyond the frontiers of the Arabian peninsula; the
latter to the subsequent imposition of religious uniformity upon subject
peoples by an Islamic power. While not at all concerned with the problems
being examined here, Becker's work provided a valuable lesson in the
typological differentiation of source materials, according to which 'le
fait coranique' might indeed qualify as 'fact' but apparently only as one
whose historical effect was (temporarily) suspended. On that particular
point there is at least room for argument, 2 but the example is none the
less instructive. I should like to propose a metaphorical extension of
Becker's dichotomy by defining Arabicization as the expression of a
centrifugal force and figure of expansion; conversely, Islamization could
be interpreted as a centripetal force and figure of contraction. Materially
the antithetical figures might be understood to represent, on the one hand,
the spread of Arabic dialects at a pace approximately consonant with that
of the Arab conquests, and on the other, the imposition of CA as a lin-
guistic/literary standard. Symbolically the contrast would be one between
natural, uninhibited diffusion and artificial, consciously directed restric-
tion. Linguistically Arabicization is characterized by a concept of language
as the most convenient means for meeting the demands of normal com-
munication (Mitteilungsbediirfnis), and Islamization by a concept of
language as an instrument of education (Bildutigsprinzip). An example of
the first was the introduction of Arabic as official language into the Umay-
yad chancery; the example of the second was, of course, the Qur'an.

Now, it has for some time been assumed that both the chancery language
in question and that of the Qur'an represent CA, and further, that both
reflect the 'poetic koine' asserted to have been the medium of pre-Islamic
poetry. Since all three may be described as forms of literary expression,
the equation is at least theoretically valid. And from the notion of con-
tinuity implicit there was born the concept of koine as descriptive prin-
ciple in the historical analysis of CA. Whether or when the koine might
have been even approximately similar to an Arabic vernacular is a problem
still awaiting solution, or reformulation. Rather more important, in my
opinion, is whether the equation is in practice justifiable. May one, in
fact, identify the language of poetry with that of scripture and, in turn,
both with that of the earliest chancery papyri ? Further, what chronological
conclusions can legitimately be drawn from the purely linguistic content
of that material ? Orthographic and morphological features common to all
three permit a degree of synchronic comparison. Lexical, syntactic, and
above all, stylistic requirements of each of the three genres might be

1 'Der Islam als Problem', 6-7. 2 See above, I pp. 43-52.

go

QURANIC STUDIES

thought to preclude a diachronic analysis. Assertions of linguistic con-
tinuity can here only be based upon non-linguistic data, and the first
century of Islam can hardly be described as 'a period fully in the light
of history'. 1

From employment of the koine concept as an explicative mechanism to
assertion of its historicity is, after all, a long and significant step. If it is to
be attempted at all, the points of departure must be an examination of
five kinds of linguistic material, each with some claim to be representative
of the earliest form of CA:

1. Poetry (pre-Islamic and early Islamic).

2. Qur'an.

3. Hadith (subsuming sira-maghdzi literature).

4. Ayyam.

5. Papyri.

With the exception of the papyri, use of these data is subject to three
caveats of general nature as well as to a number of specific reservations for
each category. First, the purpose of each may be described as essentially
educative rather than merely communicative. Literary composition with
an avowed aim must be evaluated in terms of the rhetorical schemata
consciously employed to achieve that aim, whether aesthetic, cultic,
juridical, or historical, and seldom if ever constitutes disinterested lin-
guistic evidence. 2 Second, a curious quality of simultaneity characterizes
the recensions in which this material has been preserved, as well as the
commentary and other critical literature generated by it: all appears to
have come into existence at the end of the second/eighth and beginning of
the third/ninth centuries. A concomitant homogeneity of subject-matter is
reflected in the overlapping of genres: poetry placed within a narrative
frame, prose relieved by poetic insertions both as exegesis and ornatus,
juridical and lexical problems solved by reference to scripture, scriptural
problems solved by reference to jurisprudence and lexicography, all with
approximately the same end in view: historical description of a situation
two centuries earlier. 3 Third, practical application of the linguistic data
preserved in that literature has tended to blur what might be thought a
traditionally valid distinction between prose and poetry: freely variable
syntax as a function of rhetoric. Formulation of grammatical rules from
poetry, from the style of scripture, or from ornate prose is bound to result
in a very specialized table of correct procedures, not in a generally useful
description of language. Resort to such may produce a circumscribed and
unrealistic image, which will thus become a decidedly unattainable ideal.

1 As in Rabin, 'Beginnings', 29.

2 See Wehr, Review: Fiick, 'ArtMya, 185.

3 See Spitaler, 'Arabisch', 125.

ORIGINS OF CLASSICAL ARABIC

9i

From the point of view of the Arab grammarians that may not have been
unintentional. Results are, after all, as much conditioned by method as by
material. 1

It might seem from these observations that the real value of this material
lies in its relevance to literary and scholarly activities within the community
of Islam at about the turn of the second/eighth and third/ninth centuries.
Such, at least, would be my provisional conclusion, but it is worth noting
that despite the limitations set out here and acknowledged elsewhere,
acceptance of the sources at face value continues to find adherents. 2
Now, the single exception to a very strong possibility of Nachdichtung
is the linguistic content of the chancery papyri. That has been assessed
as exhibiting CA with some slight deviation of colloquial origin, or
Middle Arabic admixture. 3 However that early chancery language may be
described, it is clearly not the language either of poetry or of scripture.
It could thus be allocated to the sphere of CA only by reference to a
standard posterior in time to the composition of the documents them-
selves. Consideration of the resultant anachronism may have led Rabin
to the hypothesis that the language of the papyri might be 'a Classical
Arabic (sicl) not yet standardized by grammarians'. 4 In practice it is
essential to distinguish between adaptations and translations from the
Byzantine epistolaris sermo and free composition in what could be called
sub-literary or 'business' Arabic, but also to observe that the chancery
language as such eventually was, and in fact had to be, elaborated
from both sources. 5 Consistent deviation from a hypothetical CA in the
language of the papyri may only be interpreted as such by reference to a
standard for which there are no extant loci probantes. Even the literary
papyri studied by Abbott exhibit the same departures from 'classical'
Arabic. 6 It might indeed be thought that these reservations contribute to
undermine the unitary concept of a 'literary koine' and, more especially,
of its development no later than the sixth century.

Lexical and syntactical disparities between the styles respectively of
poetry, scripture, and chancery documents represent functional cleavages
of a sort difficult to reconcile with the notion of a single source for CA.
It can, in fact, be argued that neither poetry nor scripture could have, or
ever did, become canons of actual linguistic usage, as opposed to sources

1 Spitaler, 'Arabisch', 126; id. Review: Bloch, Vers und Sprache, 317-18; Ullmann,
Ragazpoesie, 218-32.

2 e.g. Rabin, 'Beginnings', 21-2; the same curious and quite illogical position seems to
be that underlying Blachere's in many respects very useful Histoire.

3 See e.g. Grohmann, Einfuhrung, 103-7; Blau, Emergence, 123-32.

4 EI, s.v. 'Arabiyya: I, 564.

5 Consider the rhetorical components of classical inshd', as extrapolated from Qalqa-
shandi by Bjdrkman, Staatskanzlei, 87-92; but also the medieval European development
of cars dictamims, Curtius, Europaische Literatur, 83.

6 e.g. SALP ii; my references in BSOAS xxxi (1968) 614-16.

92

QURANIC STUDIES

of grammatical theory. 1 Assertions to the contrary exhibit theological
orientations defined to accommodate two dogmas inextricably part of the
traditional assessment of Islamic origins, namely, those of the inimitability
of the Qur'an and of the rhetorical potential of the Arabic language. 2
Both played considerable roles in the establishment of a chronological
and geographical framework for the diachronic description of CA. On the
other hand, the evolution suggested by an examination of the papyri
displays what might appear to be a normal process of refinement : from the
exigencies of administrative communication to the luxury of an elaborate
Kunstprosa in the third/ninth century. For that development we have, at
least, the requisite loci probantes: bilingual and monolingual chancery
papyri, monolingual literary papyri, the prose parts of sira, maghazi, and
ayyam, as well as the hadith literature. Save as sources of stylistic em-
bellishment (iqtibds) and lexical exotica, scripture and poetry will hardly
have affected the elaboration of a CA prose style, whether discursive or
narrative. In their earliest definitive forms these styles may be studied in
recensions of the third/ninth century, indicating thus a chronological span
which might be thought to correspond to the time-lag of two to three
centuries postulated by Becker for the separate processes of Arabiciza-
tion and Islamization. Neither religious orthodoxy nor linguistic standard
could be imposed before each had achieved canonical status.

Administrative employment of Arabic symbolized the beginnings of
acculturation. On the basis of the chancery papyri that event could be
dated to about 86/705. 3 Against the ethnic composition of the Arab domin-
ions at that time, the role of Arabic must be seen as primarily that of a
practical instrument of communication (Verkehrssprache). Its subsequent
development can be interpreted as the compound result of several non-
linguistic factors: social and economic necessity, formation of urban
centres of diffusion, and prestige of the ruling minority. Evidence of
bilingualism (substrate phenomena) and of diglossia (dialectal phenomena)
ought not to be assessed with reference to a normative CA, for which there
are no contemporary source materials, but rather with a view to the require-
ments of the community served by that sub-literary language. It is im-
portant to remember that from the beginning of that evolution to its end
the evidence is always that of a written language, whatever its pre-history
may have been. The material is thus always witness to formal and/or
formalized communication, for which emergence of a standard of excel-
lence was not only its organizing principle but also its logical conclusion. 4
The evolution so projected does not, however, require the existence of
a pre-Islamic koine, and especially not one whose linguistic description can

' See below, pp. 100-3. 2 See above, II pp. 79-83.

3 Becker, Geschichte Agyptens i, 130-1 ; Blachere, Histoire iii, 718.

4 Cf. Weinreich, Languages in Contact, 74-82, 83-110; Garbell, 'Remarks', 303-5.

ORIGINS OF CLASSICAL ARABIC

93

only be elicited from a corpus of bedouin and courtly poetry. 1 The trans-
formation of functional prose, of which the primary characteristic and
criterion is efficiency, into artistic prose will naturally reflect the impinge-
ment of a rhetorical tradition. Communicative efficiency will be gradually
superseded by a conscious striving for what is often called 'elevated
language', of which the underlying motive is not Mttteilungsbediirfnis but
Bildungsprinzip. It is then, and probably only then, that the rhetorical
embellishment associated with poetry and scripture may be seen as
operative in the formation of prose style. For the history of Classical
Arabic, however, those two sources of rhetorical schemata were defined,
preserved, and transmitted not as accessory to the basic task of communi-
cation, but as formative principles of linguistic description. The pre-
suppositions were two: (i) that eloquence and clarity were properties
exclusive to the Arabic language, and (2) that by virtue of divine election
Arabic was also a sacred language, for which change could only mean
corruption.

Any sampling of the Orientalist tradition will reveal the axiomatic
quality of Arab eloquence, 2 an impression derived presumably from the
writings of such celebrated scholars as Ibn Qutayba, Baqillani, and Suyuti,
as well as from exegesis of the tahaddi verses. 3 But the Muslim tradition,
considerably older than the dates of those authors, exhibits in its earlier
stages remarkable absence of unanimity on such questions as the presence
in Arabic of foreign lexica, the grammatical and syntactical idiosyncrasies
of poets, and the paradigmatic quality of Quranic style. 4 Predictably
dogmatic positions on these questions were taken up in the fourth/tenth
and fifth/eleventh centuries, culminating in the systematic and isagogic
works of scholars like Ibn Fans and Suyuti. 5 In a characteristic discussion
of eloquence (fasdha) as partaking of both diction and elocution, Suyuti
adduced the by that time traditional pair of assertions that the most
eloquent of men {afsah al-khalq) was the prophet Muhammad, and of the
Arabs (afsah al-'arab) Quraysh. 6 The non-linguistic nature of the argu-
ment, which might ultimately be traced to a sanctuary tradition concerning

Mecca, emerged clearly in the famous prophetical hadith i_jy*J I L>f

1 Pace Rabin, 'Beginnings', 29; cf. BSOAS xxxi (1968) 611.

2 A random selection might include e.g. Schreiner, 'Zur Geschichte', 663; Fischer,
'Usaiiid usw.', 581 n. 2; Noldeke, NBSS, 5, 22; Andrae, Person, 95-7; Blachere, Histoire
iii, 719, 725, 730; accounting thus for approximately a century of European scholarship.

3 Ibn Qutayba, Ta'tcll, 17; Baqillani, I'jdz, 24, 250; Suyuti, Itqdn ii, 270, iv, 4, 16;
id. Muzkir i, 209-10; see above, II p. 79.

4 See Kopf, 'Religious influences', 33-59.

s See Goldziher, 'Sprachgelehrsamkeit' iii, 511-52.
6 Muzhir, moo' 9: i, 184-313, esp. 209-10.

94

QURANIC STUDIES

LJ £cj J 9 ^ Jlj. 1 Suppression of claims made on behalf of other tribal
groups to die title afsah al-arab is symbolized in the account ascribed to
Farra' of how the inhabitants of cosmopolitan( !) Mecca (i.e. Quraysh) were
in a position to recognize and adopt the best ingredients from each of the
bedouin dialects in Arabia. 2 Besides drawing attention to the role of
Mecca as cultic and commercial centre, this tradition, like the ones it
eventually replaced, served to identify the northern regions of the Arabian
peninsula as the cradle of CA at a date prior to the proclamation of Islam.
Worthy of note in Suyutl's treatment of Arab eloquence is regular use of
the elatives afsah, abyan, ablagh, and a'rab in adverbial constructions with
lisdn and lugha to signify clear and effective speech. 3 Their terminological
antithesis is conveyed by a jam, glossed la yufsih, and it seems more than
likely that Quranic employment of the contrasting pair a'jami-'arabi
was intended to express just such a distinction. 4 The semantic evolution
oi~'arabi\a'rdbi is not unfamiliar: in the corresponding series deutlichj
deutenjdeutsch the formative element diutisk > tkeodiscus was, after all,
a linguistic designation prior to becoming an ethnic and ultimately a
geographical title. 5 But such an interpretation may not be adduced to
demonstrate that the locution afsah al-arab is merely a tautology; the
reference was to bedouin Arabic speech.

A natural, though perhaps not quite logical, inference from discussion
of Arab eloquence was that the language spoken by bedouin must be
identical with that of the poetry called pre-Islamic (Jahilt). Opinions of
scholars about that equation range from the almost vehement affirmation
of Noldeke across the nicely qualified acceptance of Blau to the outright
rejection of Wehr and Spitaler. 6 Even qualified assent will postulate a dual
role for the bedouin in question: (i) as referees if not arbiters in linguistic
disputes, and (2) as preservers and transmitters of Jahili poetry. Emphasis
upon the first of these roles is symbolized in Muslim tradition by the
claim of Basran grammarians to have got their linguistic information from
none but authentic dwellers in the desert: oUiJI 'J^j>. ^ Lull Ju.tl jxi,
a figure expanded to include the second role and made the object of not very
subtle caricature in the fifth/eleventh century by Ma'arri. 7 Though stories
of the unreliability, venality, and even treachery of bedouin informants

1 Cf. Mehren, Rhetorik, iao-1, but also Rabin, West- Arabian, 22-3.

2 Suyu{i, Muzhir i, 221; cf. Kahle, 'Readers', 70-1: the story was pressed into the
service of a number of distinct but related causes; for the literary effect of similar tradi-
tions see also above, I pp. 42-3, II pp. 69-70.

3 Muzhir i, 202, 200-10 and passim. 4 See below, pp. 98-9.
s Cf. Weisberger, Deutsch als Volksname, 96, 252, 278-86; Kluge-Gotze, Worterbuch,

132-3.

4 Noldeke, BSS, 4-5 vs. Wetzstein, see below, p. 249; Blau, 'Bedouins', 42-51;
Wehr, Review: Fuck, 183-4; Spitaler, Review: Fuck, 145-6.

7 See Weil, Schulen, 4 m. 1; Ma'arri, Risdlat al-ghufrdn, 168-9 on the name/sobriquet
of the mukhadrami poet A'sha.

ORIGINS OF CLASSICAL ARABIC

95

may reflect nothing more than school disputes, it would seem that these
hardly functioned as disinterested referees. 1 It might be thought that the
principle significance of the bedouin contribution, whether real or fabri-
cated, to the formation of CA lay in the grammarians' concept of i'rab: as a
term of linguistic description it may, after all, denote a self-conscious
process of bedouinization. 2 As reference to a historical event i'rab is
supported by evidence of a substantial Volkenvanderung* For our purposes
it is well to remember that the written record of transactions between
bedouin and philologist dates only from the third/ninth century, and is
thus coincident with the literary stabilization of both Quranic exegesis and
Muslim historiography.

The reputation of bedouin as custodians of Arab eloquence rests upon
the link connecting them with the creation, preservation, and transmission
of Jdhili poetry. For at least the second and third elements of that link
Noldeke's pessimistic assessment in 1864 was not materially modified by
Blachere's cautiously optimistic account written nearly a century later. 4
Such factors as variae lectiones, infinitely variable line sequence, and a
system of attribution which can only be described as irresponsible, do not
inspire confidence in the philological tradition. On the other hand, dis-
putes about the authenticity of the Jdhili poetry seem to me almost entirely
futile, so long as the evidence {pro et contra) is assessed in the light of
traditional chronology. Two examples of this approach were the essay of
Margoliouth and the retort of Braunlich: 5 acceptance of the identity of
scriptural with poetic language, of the historicity of the prophet's quarrel
with the poets, and of the allegation that poetry was from the very first
employed for scriptural exegesis (i.e. by 'Abdallah b. 'Abbas), exhibits
an unjustifiable and quite unnecessary acquiescence in the data of a nor-
mative tradition. It may be useful to distinguish between the content of the
poetry and the use made of it by philologists. Assent to a putative bedouin
environment requires analysis of both theme and imagery. Elements of
either for which only bedouin origin is conceivable are very few indeed. 6
And even those will not support unequivocal conclusions about the date of
composition of the verses in which they appear.

Such at least is the only legitimate inference from the contents of the
anthology transmitted as the legacy of B. Hudhayl, thought to have been
composed between 550 and 700. 7 Its value as historical source, for the

1 Cf. Blachere, Histoire i, 117-27.

x Cf. Fleisch, EI, s.v. I'rab; and see below, pp. 1 06-11.

3 See Caskel, 'Zur Beduinisierung Arabiens', *28~36*.

* Kenntnis der Poesie, vi-xvii; Blachere, Histoire i, 83-186, esp. 166-79.

5 'Origins', 417-49; 'Frage der Echtheit', 825-33; cf. Shahid, 'Contribution', 564 n. 3,
following Arberry and Gibb.

6 See Blachere, Histoire ii, 368-453 ; Braunlich, 'Versuch', esp. 222-38.

7 Braunlich, "Versuch', 201-11.

96

QURANIC STUDIES

origins both of CA and of Islam, must be accordingly modified. Bedouin
poetry was not so much 'pre-Islamic' as it was 'un- Islamic', a distinction
surely of some relevance to its use in solving chronological problems.
Delineation of ajahili ethos by reference to the binary opposition muruwwa
(virtus, dperrj):din (religion, law) is subject to the same reservation. 1
Goldziher's conclusions on the nature of the conflict between Islam and
paganism reflect an arbitrary chronology imposed upon rather than
elicited from his source materials. Neither muruwwa as the embodiment of
valour (hamasd), nor din as essentially asceticism (zuhd) exhibits a chrono-
logical line of demarcation between Islam and Jahiliyya. Bedouin rejection
of religious prescription represents, after all, a constant in Islamic history,
and can more profitably be interpreted as a reflex of social, economic,
even regional dichotomies, and chronologically unlimited. 2 For the bulk
oijahili poetry information is more, not less, refractory than that pertinent
to the Hudhali diwan: quite apart from the notorious— tack of scruple
attributed to transmitters (ruwat), themselves often ambitious apprentice
poets, biographical notices of Jahili authors are known to conform to a
very few stereotypes extrapolated from the imagery of whatever verse
might (fortuitously) have been ascribed to them. 3 That historiographical
technique, by means of which metaphor generated reality, was the object
of considerable parody in those passages of Ma'arri's R. Ghufran depicting
the encounters of his protagonist Ibn al-Qarih with a series of celebrated
poets. 4 Easily the most famous example of aetiological exegesis is that of the
Mu'allaqat being 'suspended' in the Ka'ba: a combination of sanctuary
tradition and witness to Arab eloquence was a temptation not easily
resisted, even though only as an afterthought. 5

It is, curiously, the religious imagery in hadari (!) poetry, historically
associated with the 'seeker of God' (hanif) in Arabia, that appears to
exhibit a link with the themes of Islamic revelation. Hirschberg's treatment
of that material illustrated its ambivalent documentary value. 6 For each
element there three distinct problems arise : relation of the verse in question
to the Judaeo-Christian tradition (whether written or oral); relation of the
verse to the Qur'an; and finally, the authenticity of the verse. But solutions
to the last tended to be subjective, and naturally based upon the traditional
chronology of Islamic origins. Rejection of an intentional Muslim forgery
of that poetry on the grounds that such would have undermined claims to
originality made on behalf of the prophet of Islam is ingenuous: both the

1 As in Goldziher, Studien i, 1-39.

* Cf. Bravmann, 'Background', 317-24, who regarded the concepts of muruwwa and
din not as antithetical but as complementary.

3 Blachere, Histoire i, 96-107, 161-6.

4 Ma'arri, Risdlat al-ghufrdn, 187 ff: 'Adiy b. Zayd; 191: Abu Dhu'ayb.

5 Blachere, Histoire i, 144-7, following Noldeke, Kenntnis, xvii-xxiii.

6 Lekren, 3-io.

ORIGINS OF CLASSICAL ARABIC

97

\ content of scripture and of the exegetical tradition presuppose close
association with Judaeo-Christian sources. On the other hand, the hypo-
thesis of widespread forgery is unnecessary: isolated occurrence (however
numerous the examples) of the schemata of revelation does not indicate
imitation of the canonical text of scripture, or even that the canon existed.
The recensions of source material for urban (Umayya b. Abi '1-Salt) and
court poetry ('Adiy b. Zayd, Nabigha Dhubyani, A'sha) are, after all, quite
late. 1 It is not entirely insignificant that precisely this corpus of poetry was
not accepted by (Muslim) philologists as evidence of Arabian fasdha. 2
But literary analysis of the schemata of revelation must take account
of the possibility at least of Arabic Vorlagen, and an Arabian tradition of
monotheism might be thought, from the point of view of later Islamic
orthodoxy, to have provided more appropriate reference than Arabic
versions of Jewish and Christian materials. That such reservations were
not, however, characteristic of the earliest Muslim exegesis is clear from
examination of the haggadic literature. 3

Whatever may have been the original motives for collecting and record-
ing the ancient poetry of the Arabs, 4 the earliest evidence of such activity
belongs, not unexpectedly, to the third/ninth century and the work of the
classical philologists. The manner in which this material was manipulated
by its collectors to support almost any argument appears never to have
been very successfully concealed. The procedure, moreover, was common
to all fields of scholarly activity: e.g. the early dating of a verse ascribed
to the mukhadrami poet Nabigha Ja'di in order to provide a pre-Islamic
proof text for a common Quranic construction (finite verb form preceded
by direct object), 5 Mubarrad's admitted invention of a Jahili verse as gloss
to a lexical item in hadith, 6 and Abu 'Amr b. 'Ala's candid admission that
save for a single verse of 'Amr b. Kulthum, knowledge of Yawm Khazaz
would have been lost to posterity. 7 The three examples share at least one
common motive: recognition of pre-Islamic poetry as authority in lin-
guistic matters, even where such contained non-linguistic implications.
Also common to all three is another, perhaps equally significant feature :
Ibn Qutayba, who adduced the verse of Nabigha to explain/justify Quranic
syntax, lived at the end of the third/ninth century, as did Mubarrad;
Abu 'Amr, of whom no written works were preserved, lived in the second
half of the second/eighth century, but this particular dictum was alluded

1 Hirschberg, Lehren, 14, 32-40.

2 Hirschberg, Lehren, 25-6, n. 3 ; see below, pp. 105-6.

3 See below, IV pp. 122-48.

* A not very convincing enumeration in Blachere, Histoire i, 94-5.

5 Spitaler, Review: Bloch, 320. 6 Margoliouth, 'Origins', 431.

7 Ibn 'Abd Rabbih, Al-'Iqd al-farid iii, 106-7; the problematic character of this
particular one of the ayydm al-'arab may be guessed from its inclusion in Jahiz, Kitdb
al-Tarbi', 101.

4339C76 E

9»

QURANIC STUDIES

to only in Jahiz (third/ninth century) and explicitly stated in Ibn 'Abd
Rabbin (fourth/tenth century). Now, that pre-Islamic poetry should have
achieved a kind of status as linguistic canon some time in the third/ninth
century may provoke no quarrel. That it had achieved any such status
earlier must, I think, be demonstrated. The fact that it had not, in one field
at least, can be shown : the absence of poetic shawdMd in the earliest form
of scriptural exegesis might be thought to indicate that appeal to the
authority of Jahili (and other) poetry was not standard practice before the
third/ninth century. 1 Assertions to the contrary may be understood as
witness to the extraordinary influence exercised by the concept of fasahat
al-jahiliyya.

The utility of that concept is nowhere more apparent than in the inter-
pretation of what otherwise might be held comparatively neutral references

to language in the text of the Qur'an. That Q. 14: 4 Jj^j ^ LL.jt Lj

Juy qLA) VI could be made to bear diametrically opposed meanings

has been noticed. 2 That generous view was, however, not shared by all
exegetes, and Zamakhshari was quite explicit on the point. 3 His cross-
reference there to Q. 41: 44 aJU ol^ki VjJ IjJUJ U^f Ul^s oUL^. jjj
is significant because in that verse he interpreted fussilat as

buyyinat (clarified) and the contrast between a'jami and 'arabias the absence
and presence, respectively, of clarity. 4 The polarity reflects, of course, the
terminology of arguments about (Arab) eloquence, 5 but for Q. 41: 44
produced, illogically, the meaning: is it an unclear (qur'an) and an Arab
(messenger)? The alternative: is it an unclear (qur'an) and (yet) an elo-
quent (messenger), is not absolutely excluded, but Zamakhshari's addi-
tional observation that a'jami might refer to barbarians (ummat al-'ajam)
makes the former interpretation rather more likely. Evidence of the
transition from a' jam as inarticulate to a'jatnj'ajam as barbarian/non-Arab
emerges from the literary references to mawali in Islam assembled by
Goldziher, especially where inferior social status was the consequence of
both foreign blood and defective speech, e.g. a'jamu timtimiyyun, Ziyad
al-a'jam, and alhan al-mawali. 6 1 am inclined to interpret a' jam as originally
a term of linguistic description.

For a related verse, Q. 16: 103 jU j^i «uU> U! djijk fX** JGdj
,2^> (^jj* jl—J I .aaj Oj-J^bIj t£jJI Zamakhshari defined a'jami

as ineloquent (ghayr bayyin) and 'arabi mubin as clear (and) eloquent (dhu
bay an wa-fasaha). 7 It may be that collocation here of mubin (clear/obvious)

1 See below, IV pp. 216-18. * See above, II p. 53.

3 Kaskskdfii, 538-9 ad loc. 4 Kashshdfiv, 202. 5 See above, p. 93.

6 Studieni, 101-46, esp. 103 n. 2, 118 n. 2, 120.

7 Kashshdf ii, 635.

ORIGINS OF CLASSICAL ARABIC

99

with 'arabi facilitated this interpretation, but the same collocation in Q.
26: 194-5 ^js. o\Jh ^jJuJI ^ Oj5cJ dLii Jx. provoked the

comment that 'arabi must refer to the Arabic language because the heart of
a prophet could only be affected by hearing his mother-tongue I 1 Similarly,
collocation of hukm with 'arabi in Q. 13: 37 produced the explanation: an

Arabic decision, translated into the language of the Arabs i»5C>.

). 2 Collocation of qur'dn with 'arabi (Q. 12: 2, 20: 113,

39: 28, 41 : 3, 42 : 7, 43 : 3) did not invariably provoke the same response,
and indeed for Q. 39 : 28 the apposition ghayra dhi 'vwajin (without dis-
tortion) might be thought anyway to have precluded such. Most occur-
rences of lisdn in the text of scripture exhibit the use of 'tongue' as vocal
organ rather than as language, and might appear to reflect the speech
difficulties associated with the calling of Moses (cf. Exodus 4: 10-17).

The locutions (jLJ ^ i"Juic Jis-lj (Q. 20: 27) and UU JL (28: 34J

refer of course to Moses, and it is worth recalling that the verses following
upon Q. 14: 4 also take up the Mosaic narrative, in a manner indeed
which suggests that the original function of verses 1-4 could have been to
introduce that narrative. A similar instance was noted for the isra' verse

Q. 17: i. 3 The imagery of Q. 19: 97 and 44: 58 dJLL fiU^j UU as well

as of 54: 17, 22, 32, 40 (serving as refrain) o\yi\ li^j Juij is more than

merely reminiscent of Exodus 4:15 T B ~ D57 riTlS "DUO, and could support
the hypothesis that linguistic allusions in the Qur'an are not to the Arabic
language but rather, to the task of prophetical communication. 4 That such
eventually involved identification of the instrument by means of which the
word of God was transmitted may seem not unreasonable, but the evi-
dence of scripture itself yielded assistance only under duress.

The concept of lingua sacra, like that of the mythopoeic fasahat al-
jahiliyya, belongs to a view of language as criterion of culture. The primary
function of language so considered is not communication but edification.
Its semantic spectrum is selective, its syntax stereotype and rhetorical,
its style paraenetic. Whether specialized form is a consequence of special-
ized function, or vice versa, may be impossible to determine, at least as a
generally valid proposition. What does seem clear is that linguistic data of
that kind cannot be very profitably examined from the standpoint of 'the

1 Kashshdfiii, 334-5.

2 Kashshdf ii, 533; cf. also iv, 301 ad Q. 46: 12 and lisdn 'arabi.

3 See above, II pp. 67-8.

4 The traditional view is variously expressed in Horovitz, Uniersuckungen, 75 ; Torrey,
Foundation, 43—4; Obermann, 'Agada', 47-8; Widengren, Apostle of God, 15 1-2.

IOO

QURANIC STUDIES

ordinary day-to-day requirements of a normal speech community'. 1
And yet, the derivation of normative grammar from scripture and poetry,
if it ever meant more than merely theoretical accommodation of non-
linguistic (e.g. theological) postulates, supposes just such a point of view.
As a genre of narrative prose the scriptural style was examined by Auer-
bach in an essay designed to emphasize the contrast between Homer and
the Elohist. 2 Of the nine factors adduced by the author as characteristic
of Old Testament narrative, one in particular deserves mention here: the
need for exegesis (Deutungsbediirftigkeit). This is not merely to say that
the content of scripture is enhanced by commentary, or that it may be
made to bear any number of (complementary and/or contradictory) inter-
pretations, but that the scriptural style is itself incomplete without com-
mentary. Reasons for that condition were partly syntactic (Abgerissenheit,
Stilmischung), partly rhetorical (Vielschichtigkeit, Hintergriindigkeit) . The
analysis can be extended to include at least one additional element, namely
the symbolic quality inherent in scriptural diction. That quality may be
regarded as one fundamentally semantic, enhanced by the crystallization
of imagery which achieves an existence and application independent of
its constituent elements, a distinction which might be covered by the
contrast information vs. reference. 3 An example from Biblical literature
was the Aramaic locution tMK For Muslim scripture those elements
analysed here as schemata of revelation, appropriate to the theodicy and
to prophetology, exhibit the metamorphosis of originally neutral notation
conveying 'information' into significantly charged 'references'. The under-
lying change may be described as rhetorical, the references themselves as
symbols (Stichzvorte) both requiring and producing exegesis. The inter-
pretation thus generated spans a range limited to the spectrum of allusion
contained in the imagery. For the Quranic revelation that range was pre-
selected by the more or less unqualified employment of Biblical schemata.
Analysis of the limits of interpretation is the subject of the following
chapter; the question to be examined here is whether this particular
example of lingua sacra could, or did, provide genuinely useful data for the
grammar of Classical Arabic.

To what extent establishment and transmission of the text of scripture
involved its grammatical analysis is a problem to which tentative solutions
often reflect preconceptions elicited from non-linguistic data. For the
Hebrew Masorah Bacher's view was that little, if any, grammatical
description resulted from the activity of the Masoretes and then, as it
were, only accidentally. 5 Later studies of the same phenomena reveal a

1 Ullendorff, 'Is Biblical Hebrew a language?', 241. 1 Mimesis, 5-37.

3 Barr, Comparative Philology, 118, 291-3; id. Semantics, 206-62; Vermes, Scripture,
11-66: on exegetical symbolism; see below, IV pp. 210—12.

4 See Vermes, 'Jewish Aramaic'.

5 'Die Anfange der hebraischen Grammatik', 8.

ORIGINS OF CLASSICAL ARABIC

IOI

rather more generous assessment of Masoretic criteria, even if these had
not produced descriptive terminology. 1 For the document of Islamic
revelation acceptance of the 'Uthmanic recension traditions has entailed
assent to a period of from 150 to 200 years between textual stabilization of
the Qur'an and analysis of its contents in the formulation of Arabic
grammar. The implication must be that the text of scripture, like those of
pre-Islamic poetry, was faithfully transmitted and intelligently read/
recited and heard for a very long time indeed, without once provoking
the questions about its meaning and its form with which the literature of
the third/ninth century is filled. Logic alone might preclude serious
consideration of this version of Islamic history. Examination, moreover, of
the Quranic exegesis which I have called masoretic suggests that both the
document of revelation and the corpus of pre-Islamic poetry were being
there assembled, juxtaposed, and studied for the first time. 2

Systematic collation of loci probantes from both sources, such as informs
Zamakhshari's Mufassal (sixth/twelfth century), represents the final stage
of a long period of synthesis. The theological, as opposed to linguistic,
nature of that dual source of grammar has been noticed. 3 It may further be
observed that Zamakhshari's concern with grammar in the most restricted
sense of that term (i.e. morphology and juncture) permitted omission of the
broader issues of syntax and even context, in consequence of which his
loci probantes are stylistically neutral. As descriptive material they are also
linguistically indifferent, consisting almost exclusively of deliberate and
formalized utterances postulated as paradigms, from which deviation could
only be regarded as corruption. The method is familiar, its results pre-
dictable. That these could, and occasionally did, provoke opposition seems
clear from the formula attributed to Basran grammarians in their rejection
of the Kufan practice of drawing grammatical analogies from scripture:

Jjyjj VI J-ta l&J^ ^yi V iSy-s ■> J***-* Such did not, of course,
prevent the Basrans' adducing scripture where this supported their own
arguments. 5 Unlike Zamakhshari's normative grammar, Ibn Anbari's de-
scription of grammatical dispute (also sixth/twelfth century) includes con-
siderable discussion of syntax, though preponderantly of configurations
generated by the use of tmesis, hyperbaton, chiasmus, etc. (taqdim wa-
ta'kkir), for which the styles of scripture and poetry might be thought
appropriate arbiters. Neither work is, thus, free of concern with preciosity,
whether grammatical or stylistic. In both, rhetorical ideals pass for linguistic

1 Cf. Wiirthwein, Text, 11-22; Corr£, 'Phonemic problems', 59-66; Gerhardsson,
Memory, 43-55, esp. 52 n. 3 on the work of Xahle; Barr, Comparative Philology, 202-7,
and 214-17 on Kahle.

2 See below, IV pp. 202-27.

3 Cf. also Vollers, Review: Noldeke, 126-7; Kopf, 'Religious influences', 46-50.
♦ Ibn Anbari, Al-Insdf, 137, cf. also 264, 392; Weil, Schulen, So n. 1.

5 e.g. Ibn Anbari, Al-Insdf, 47, 53, 61, 7 and passim.

102

QURANIC STUDIES

description and the concept of grammar as a balance between analogy
and anomaly with reference to an acceptable prose style is absent.

It is, however, necessary, even within this restricted framework, to ask
whether the document of revelation ever became in any practical sense a
model of linguistic usage. Probably the only recorded instance of agree-
ment between Vollers and Noldeke was about the character of Quranic
syntax. 1 But while Noldeke rightly concluded that such a collection of
anacolutha must remain sui generis, Vollers shifted his ground just slightly,
in order to bring into existence the hypothesis associated with his name
since 1905: that the original (vernacular) text of the Qur'an had been
refashioned to meet the linguistic standard represented by pre-Islamic
poetry, and so to produce what is known as CA ('arabiyya). 2 That view,
which has found more or less unanimous dissent for reasons both sound
and not so sound, involved a change of emphasis from syntax to accidence,
more specifically to the properties of inflexion. The argument revealed a
seductive, if in the event erroneous, logic, since it was only in the realm of
inflective phenomena that Quranic usage might be seen to intersect that of
Jahili poetry. The basic error lay in Vollers's adherence to an arbitrary and
Active chronology, 3 though that may have been less important than his
contention that the refashioned language of scripture could be identified
as the CA of the Arabic grammarians. Neither from the point of view of
lexicon nor from that of syntax could the claim be justified.

An intrinsic feature of lingua sacra is its 'elevated style'. While there is
undoubtedly something to be said for interpreting that characteristic as the
impingement of poetic syntax upon prose, it may also be something more,
or less, or quite different. 4 If it is true that Bloch's investigations resulted in
a statement which tended to justify extrapolation of normative grammar
from poetic texts, that was probably because his comparative material was
drawn from what can most accurately be described as Kunstprosa* His
virtual omission of Quranic shmoafdd, as well as of other examples of
rhymed prose, may thus be seen as arbitrary, since their inclusion could
hardly have altered his findings. On the other hand, the elevated style
(gehobene SprackejVerheissungsstit) associated with Kunstprosa need not
have had its origins in poetry. The transformation of sub-literary, even
vernacular, language into composition of sustained elevated style may be
effected by nothing more than resort to an established rhetorical tradition. 6
An example was the creation of medieval Hebrew poetry, which, in
addition to metrical features adapted from Arabic verse, depended for its

1 Vollers, op. cit. 127; Noldeke, NBSS, 22-3.

2 Volkssprache, 175-85.

3 See above, p. 86.

* See Bloch, Vers und Sprache, 30-1, 39; cf. Spitaler, Review, 320.

5 Vers und Sprache, 154-5: 'tabellarische Obersicht'.

6 Cf. Barr, Comparative Philology, 227-32: on 'restorations'; id. Semantics, 263-87.

ORIGINS OF CLASSICAL ARABIC

103

aesthetic effect upon the incorporation of Biblical imagery, and hence
Biblical syntax. 1 Such did not, of course, limit adaptation(s) from the
Arabic poetic tradition, though the re-creation of these in Hebrew in-
volved yet a further step in the mimetic process. But introduction of
Biblical imagery, whether in Hebrew poetry or in Arabic scripture, may be
regarded as the agent of an altered status for the language employed, and a
reflex of style as Bildungsprinzip. 2 In the lingua sacra itself the principle of
edification often took the form of linguistic archaism. A conscious resusci-
tation of past glory by recourse to language is symbolized for seventh-
century Israel in the figure of Josiah and Deuteronomy (2 Kings 22-3). 3
In that and other examples of archaizing tendencies in literature the pur-
pose was invariably paideutic (Gemeindebildung)* and often conceived as a
panacea in conditions of social fragmentation. The linguistic tradition to
which reformers and prophets, as well as poets, turn may be ancient. What
it must be, is other than the current usus loquendi, and instances of that
kind of archaization are abundant. 5

The allegation of antiquity might seem a valuable corollary to the role of
language in communal reformation. In Arabic philology the formative
principle of language as sacred was applied also to that problem. Exegesis

of Q. 2 : 3 1 l^Lt" e.U-.Vl ^ i\ jjiftj ranged from the straightforward assertion

that language was a divine creation, leaving the Quranic verse diametrically
opposed to Genesis 2: 19-20, to the modified proposal that God enabled

Adam to bestow the names <j\ ^s- fiT jJil, bringing thus the

Quranic and Biblical notions somewhat closer. 6 A concept of purely rational
revolution (istildh) appears to have been displaced, or modified beyond
recognition, by a number of pseudo-evolutionary theories linking the gift
of pure speech with that of revelation (tazvqif), historically related to the
period encompassed by Ishmael and Muhammad. 7 Implicit admission that
pre-Ishmaelite Arabic (sic) was impure tended to be muddled by the
assertion that the word of God (kalam allah) was Arabic. 8 But the marginal
development thus admitted had to cease with the revelation to Muhammad :
cJ J». oJuo ^ i*J ^i. 9 The natural but not quite necessary transition

1 See Goldziher, 'Bemerkungen zur neuhebraischen Poesie', 719-36; Mirsky, 'Biblical
variants', 159—63; Gertner, 'Medieval Hebrew writing', 163-93.

1 See Rabin, 'Qumran Hebrew', 144-61; Koch, Growth, 106-7; Seeligmann, 'Mi-
draschexegese', 163.

3 Cf. von Rad, Theology i, 77, and 81: on Lamentations 5: 21.

4 See above, I pp. 51-2; II pp. 77-8, 82-4.

5 Rabin, 'Qumran Hebrew', 159; id. West- Arabian, 17, references to Cantineau,
Doughty, and Socin.

6 Suyuti, Muzhir i, 8-1 1; cf. Geiger, Was hat Mohammed, 96-7; Kopf, 'Religious
influences', 55-9.

7 Suyuti, Muzhir i, 27-35. 8 See above, II pp. 77, 83-4.
• Suyuti, Muzhir i, 9.

io4

QURANIC STUDIES

from word of God to language of God was not, of course, confined to
Islam. Vaguely contradictory traditions achieved a remarkably comfortable
co-existence : on the one hand, the seventy/seventy-two languages in which
the Torah had been revealed/interpreted was reflected in the Pentecostal
gift of tongues and in the Muslim version of the confusion at Babel; 1 on
the other, the subsumption of indifferent Biblical designations of the
Hebrew language under the Rabbinic concept of leshon ha-qodesh, reflected
in Patristic assessments of both Septuagint Greek and Vulgate Latin, 2 and
in the Muslim dogma of tawqtf. Scriptural references to the language of
God, consisting for the most part either of metaphorical extensions of
the 'word of God' or of allusions to the organ(s) of speech, could prove
refractory. A passage like Ezekiel 3 : 5 ]wb HBP •>)?»» as-Vx X 1 ? "O
VXHF rva Vx ni?V MIX may be thought typologically akin to Q. 14: 4
0UJ4 VI Jj-«j ^ UL«jf Uj: both exhibit the essentially ethnic
orientation of the prophetical mission. 3 Neither verse restricted in the least
theological inferences that identified the vehicle of revelation with the
language of God.

If that language was Arabic internal variations had none the less to be
isolated and eventually eliminated. The kind allocated to the period before
Ishmael, like the varieties associated with Qahtan and Jurhum, could be,
and were, superseded by the fact of the Quranic revelation. 4 The at least
theoretical identification of linguistic with confessional community which
resulted from that argument (in so far as it ever achieved unanimous
resolution) yielded a standard by which dialectical differences (lughat)
could be equated with heterodoxy and worse. A succinct statement of that
procedure appears, unexpectedly, in Suyutl's discussion of the imala and
related phenomena, and in the form of a prophetical hadith: q\ jA\ I jj J\
^ot53! J.aIj jj^iJI J.&I cj\j*>]j l^Jlj^lj (^jjxll Ojxb (Recite

the Qur'an with the rhythmic embellishment of the Arabs, and avoid that
of sinners, Jews, and Christians). 5 Determination of the kind and degree
of tendentiousness in that utterance will depend upon which of a multi-
plicity of interpretations is finally selected : I cannot but suspect that the
hendiadys ( ?) luhiin wa-aswdt is a reference to i'rdb. 6 What does seem clear
is the delimitation of yet a further dialect cleavage, here based on a con-
fessional distinction: Muslims contrasted with infidels. 7 A version of the
same hadith, traced to the authority of Abu Hurayra and reported by

Tustari is more explicit: Vj U^JJ US& jJ. ^ ejyJI dy^s o[^JI

1 Cf. Jeffery, 'Scripture', 130; Geiger, op. cit. 116; Suyuti, op. cit. i, 32.

2 BergstrSsser, Hebraische Grammatik i, 9-10; Eissfeldt, Einleitung, 745-9; and see
above, II pp. 81-2.

3 See above, pp. 98-9; and II, pp. 53-4. 4 Suyuti, Muzhir i, 30-5.
5 Itqdn i, 255. 6 See below, pp. 110-11. ' See above, p. 87.

ORIGINS OF CLASSICAL ARABIC ioj
ftUu'Vl ^yutj JU AJJIj t!_yt.Vl JjfcTj ^Jlj (jJU^t J.a! OjpeL ejjju
oluill ^-ly j^jIj^I aJ Oy^-ji cS-Ajo jLu, JJlj uiiScJI

jjliUJ) diiJjl ^uUI c_JlS3I Jb*bi iJjXL* Jlc-^b There the

stigma was related specifically to the practice of synagogue and church
(i.e. cantillation), which Tustari in his commentary ascribed to 'diabolical

possession' of the kind characteristic of the Jahili poets: ^^Ju] dJU f js.
LIaUJI cl yJi, cjjlS L5~ j^^i*. 1 Now, reliable evidence of such dia-
lect cleavage, apart from assertions derived from scattered and ambiguous
testimony, hardly exists. 2 In his study of modern and contemporary
materials exhibiting this kind of phenomenon, Blanc was sceptical of those
few reports of earlier distinctions, and suggested that the later ( !) major
differentiation was probably a consequence of the bedouinization of
Muslim sedentary dialects. 3

From the prophetical tradition retailed by Tustari and SuyutI it is, of
course, impossible to draw specifically linguistic conclusions about varieties
of spoken Arabic, either during the lifetime of the Arabian prophet or at
whatever period this particular caveat may have been uttered. The state-
ment is theological, not philological, but could be of some value in deter-
mining the function of irab and, further, the role of CA in the historical
description of Islamic origins. Similarly, the story of 'Abdallah b. 'Atik's
dispatch by the prophet on an especially delicate mission 'because he could

talk like the Jews' (2b ij U J^jj, jlS" Ai ^ ) maybe understood as the historical

projection of a polemical, and very likely much later, impulse : communal
discrimination on the basis of an alleged difference in linguistic usage. 4
Severe comment on the 'Christian dialect' ('ibddiyya) of the Jahili poet
'Adly b. Zayd might be thought to reflect the same impulse, here moti-
vated by the Islamic appropriation of Jahili poetry as its rightful legacy. 5
Historical conclusions about the Christian origins of what became CA,
derived from evidence of this type, are perhaps ingenuous. 6 On the other
hand, the inverse conclusions invariably drawn from the data of 'pre-
Islamic' poetry written by Jews are simplistic; that material is subject to
the caveats applicable to the entire corpus of Jahili poetry (comprising not
only Jewish, but also Christian and hanif elements), the linguistic value of

1 Tafsir, 8-9.

2 e.g. Goldziher, 'Literatur der muhammedanischen Mystik', 766 n. 5; id. 'Me-
langes judeo-arabes', 14 n. 4; Lammens, Ardbie, 81 ; Fuck, 'Arabiya, 88; and in general,
Torrey, Foundation, 17, 47—53.

3 Communal dialects in Baghdad, 167-71, esp. 202 n. 180.

4 WaqidI, Kitab al-Maghdzi, 392; cf. Fuck, 'Arabiya, 88 n. 1; and BSOAS xxxi
(1968) 149.

5 See the caricature of those attitudes in Ma'arri, Risdlat al-Ghufrdn, 192-3.

6 e.g. Wellhausen, Reste, 232-3 ; Rabin, 'Beginnings', 27-8.

io6

QURANIC STUDIES

which cannot be separated from its theological significance. 1 The judge-
ment of Flick, that the Christian poet Akhtal could never have been recog-
nized as 'classical' had his language (sic: langue com ante) been different
from that of non-Christians, may be cited as characteristic of that kind of
argument. 3

The later evidence of Middle Arabic, found largely in texts of Jewish
and Christian provenance, belongs of course to a different sphere, and
attests to a dichotomy between spoken and written Arabic even within the
Muslim community. 3 It is none the less not impossible that the linguistic
situation described as Middle Arabic and extensively analysed in Blau's
studies was typical also of the earlier period, from which the emergence of
CA might represent a substantial deviation. 4 Against such a background
the prophetical hadith adduced by Tustari and Suyuti could acquire
added meaning: Muslims were exhorted to use a language other than that
associated with non-Muslims. Now, it may be that the language recom-
mended was also the language of God, and of Quraysh, and of the Arabian
prophet (for Quranic recitation or otherwise), but in this particular context
it was the emblem of a religious community, serving as a badge of dis-
tinction. The functional value of that language could, indeed, have
figured more than just marginally in its historical description.

Those elements of Classical Arabic which have been of most use for
studies in comparative Semitic philology are its inflective properties. From
the orthography of CA it is difficult but possible to discern specific syn-
tactic values for a range of phenomena traditionally accepted as evidence
of the preservation in CA of Proto-Semitic case and mood. 5 To infer from
this evidence that CA exhibits an archaic form of Semitic language, or that
inflexion is indispensable for its communicative efficacy, is in my opinion
unjustified. That interpretation rests, as I have in several contexts sug-
gested, upon a chronological and geographical framework that is nebulous
and anyway unverifiable. A typical example of that kind of inference is the
description of CA orthography as uniformly 'pausal' proposed by Noldeke,
a consequence presumably of his understanding of the Quranic evidence. 6
More recently W. Fischer suggested that several instances of so-called
'pausal' forms which occur in the Qur'an might, rather, be interpreted as

1 See Noldeke, Kenntnis der Pcesie, 52-86.

2 'Arabiya, 88.

3 See Blau, Emergence, 1— 18.

♦ Cf. my remarks on Blau, Grammar of Christian Arabic, in BSOAS xxxi (1968)
610-11.

5 GVG i, 459-66, 554-9; Moscati, Comparative Grammar, 94-6, 134-6.

6 BSS, 7; id. NBSS, 6: 'Aller Anfeng ist eben schwer', etc.; also GdQ iii, 27;
Birkeland, Pausalformen, 10.

ORIGINS OF CLASSICAL ARABIC 107

exhibiting the syllabic structure characteristic of modern Arabic. 1 Irregu-
larity of Quranic rhyme makes it refractory material upon which to
construct phonetic analogies, 2 but Fischer's proposals relating to the
behaviour of td marbuta (i.e. exhibiting -d->ahjst. const, -at), especially in
Sura 80, make more sense of that phenomenon than does the traditional
description in terms of an opposition at: ah J On the other hand, Fischer's
investigations caused him to postulate an early date for the onset of Middle
Arabic (i.e. 'neuarabisck') morphological structures,* rather than a later
date for establishment of the Quranic text. His reasons were the traditional
ones, and turn upon the inflective properties of Quranic Arabic.

While it may be true that the text of the Qur'an displays a linguistic
situation in which syntactic values correspond for the most part to ortho-
graphic convention, 5 that fact does not of itself yield information towards
dating establishment of the text. Moreover, the instances in which the
consonantal text does not indicate syntactic relations far outnumber those
in which it does. That such may be regarded as a fortuitous shortcoming of
the Arabic alphabet cannot diminish in any way the importance of observ-
ing that the missing orthographic detail was supplied by resort to textual
exegesis. The procedure to that end, as well as the substance itself, was
called i'rdb; the exegetical activity was the work of the Muslim masoretes
and belongs to the third/ninth century. 6 According to a long and popular
tradition, of which the most recent exponent was Fuck, parts of the Qur'an
could not be understood without i'rdb. 7 Assertions of that sort date back at
least to Ibn Qutayba (d. 276/889), whose discussion at this point reveals a
characteristic lack of concern for stress, pause, context, gesture, and other
rhetorical factors normally the accompaniment of human utterance. Thus,

the putative contrasts ^1 JJlS JJli \lst> and J^j V/Jj£> ^

^jJI Aju were claimed to contain the same inherent ambivalence

as Q. 36: 76 OjJUj L,j dj^*~i ^ |J^' tff/tfl dJj^cj ^i. 8 Of that

1 'Silbenstruktur', esp. 45-53, 53-60; vaguely anticipated by A. Fischer, 'Arab, aysh',
816.

2 See below, pp. 116-17.

3 'Silbenstruktur', 57-8 ; I am unable to follow the objection of Blau to this argument
on the grounds that because the accusative was 'too different' it was adapted to the
nominative/genitive for nouns of this category, see 'Linguistic setting', 11; id. 'L' Ap-
parition', 199 n. 3; id. Pseudo-corrections, 57 n. 14; the contrast thamdnya\thaman, cf.
id. Christian Arabic, 57-8, is not in my opinion functionally analogous to the gram-
matical differentiation of nouns ending in td marbuta; cf. also Birkeland, Pausalformen,
96-8, esp. 97 n. 1, rejected by Blau.

4 'Silbenstruktur', 59-60, but see Blau's cogent objections to that designation in the
references cited in preceding note.

s See Noldeke, BSS, 1-4; id. NBSS, 1-5.

6 See above, pp. 95, 105; and below, IV pp. 219-27.

7 'Arabiya, 3.

8 Ta'vnl, 11-12.

io8 QURANIC STUDIES

reasoning Zamakhshari made curt dismissal, explaining the non-canonical
anna as an elliptical form of li-anna (JJudl (J*) -1 Similarly,

the passages adduced by Fiick, i.e. Q. 2: 124 <dj ^1^1 ,Jal ilj; Q. 9: 3
{j£~jLA\ & <£ji -Oil jT; and Q. 35: 28 *.>L& ^ XUl ^ Li J
iLLJI could be, and were, analysed to show that any and all readings
produced acceptable, and often identical, meanings. 2 A further impression
of the manner in which this particular problem was approached by the
exegetes may be gained from the subdivision of Suyutl's chapter on
Quranic i'rdb in which are set out twenty-five verses that could be read

with all three case inflexions (jT^i) I <Jj^=» ^* dJ&Jb ijj U). 3 Even the
several much-discussed instances of orthographic deviation from the
grammatical standard of inflexion, e.g. Q. 2: 177, 4: 162,5: 69,20: 63,
could be accommodated by refining the boundaries between casus rectus
and casus obliquus.* Of these the locus classicus was Q. 20: 63 jllas jl
jl ^LJ, to the various solutions of which a good share of masoretic

method may be traced. 5 The traditional crux was the syntactic function
of the accusative/adverbial case, interpreted to cover a quite extraordinary
range of phenomena. For the Quranic text these included ambivalent value

for the particle iyyd, represented by the variant Xj£ iJbl in Q. 1: 4 ijbl
JuJ; 6 by an optional nominative/accusative in constructions based on
ammd, e.g. in Q. 41: 17 ^UjJi^i ij^J/Sj^J \J; 7 and by considerable
latitude in the formation of object pronominal suffixes with verbs termi-
nating in Unjuna, e.g. Q. 15: 54 Oj I ^ and 39: 64 jJ6\ Ji
OjIaUJI Ji^tl ^j^ti' I 4Jl5l. 8 The arbitrary, if not irresponsible,

nature of this treatment was not limited to the grammar of scripture, and
in other contexts provoked caustic observations on the activities of
grammarians, e.g. in the celebrated mas'ala zunburiyya, 9 and in the
reaction of men like 'Adiy b. Zayd, Nabigha Ja'di, and Farazdaq to

1 Kashshaf iy, 29 ad Q. 36: 76.

2 e.g. Kashshaf i, 183-4, "» 245, iii, 610-1, respectively; cf. Spitaler, Review: Fiick,
147 nn. 19,22; Wehr, Review: Fiick, 181; Fuck's fourth example, Q. 4: 8, exhibiting
orthographical evidence of case, was omitted from the French edition.

3 Itqdn, now' 41 : ii, 277-80.

4 Cf. Vollers, Volkssprache, 163, including further examples derived from variant
readings.

5 See GdQ iii, 2-6 ; Wansbrough, 'Periphrastic exegesis', 264 ; and below, IV pp. 222-4.

6 Ibn Hisham, Mugkni 'l-labib i, 96: ascribed to Hasan Basri; cf. Suyuti, Itqdn i, 216:
anonymous.

7 Kofler, 'Reste', 255-6 (Part 3).

* Wansbrough, 'Periphrastic exegesis', 252 no. 27; Vollers, Review: Noldeke, 134.

• Ibn Anbari, Al-Insdf no. 99 (pp. 702-6); cf. Fischer, 'Zunburija', 150-6; Blau,
'Bedouins', 42-51.

ORIGINS OF CLASSICAL ARABIC

109

philological interpretations of their poetry. 1 Anecdotes of that sort,
whether or not apocryphal, might be thought to cast some doubt on the
traditional description of CA.

It might also seem that the primary function of i'rdb was not, or at least
not exclusively, grammatical. If, indeed, syntactic relations could be
adequately expressed by insertion of i'rdb, it would hardly appear, even
from the few examples given here, that they could not otherwise be
expressed. Noldeke's much-cited piece justificative demonstrated that the
constituents of i'rdb were not unique in Semitic philology, but neither that
their grammatical values were constant nor that they were intrinsically
essential to the expression of syntax in CA. 2 Defence of the authenticity
(sic), as opposed to the function, of i'rdb had already been undertaken by
Derenbourg in a transparent argumentum ad hominem.* A distinction
between form and function was hardly adumbrated in the works of
Muslim philologists. Suyutl's references, as well as his own observations,
stressed the syntactic value of i'rdb for both-seripture and profane litera-
ture. 4 From his treatment of the subject, however, one impression in
particular emerges, namely, that insertion of the i'rdb into any given
locution required first knowing what it meant. 5 Not related perhaps to this
paradoxical formulation, but none the less relevant, is Suyutl's distinction
between the 'i'rdb of the grammarians' and the 'i'rdb whose recitation in the

Qur'an will be rewarded in heaven' Q^Jj A^Ul! <jU-» A3j*-> Jul il^JI

«»_>l Vj cfrly c .t J tXii £*. 6 Vollers interpreted this passage as
question-begging, and understood the i'rdb denned there to refer to 'die
stilistischen und rhetorischen Feinheiten (die aus dem grammatischen I'rab
gefolgert werden)'. 7 Whatever that may mean, it was Vollers himself whose
studies provided support for an earlier view that the function of i'rdb was
essentially that of rhythmic ornatus.* The point had been made by
Wetzstein, with reference to bedouin practice, that poetry was intended
to be sung and hence required an abundance of vowels: 'Alle diese
Bestimmungen iiber den Wegfall oder die Verkummerung der kurzen
Vocale gelten fur die Umgangssprache, nicht fur die Poesie. Das Gedicht

1 Ma'arri, Risdlat al-Ghufrdn, 183 (for 'Adiy's allegation that hamz was an Islamic
innovation), 202-3; c f- Noldeke, Kenntnis der Poesie, 31-2.

2 BSS, 1-10; cf. Corriente, 'Functional yield', 20-50; and Blau, 'Synthetic character',
29-38.

3 'Quelques remarques', 210-11.

4 Itqdn, naw' 41 : ii, 260-80; id. Muzhir i, 327-8; the preponderant role of syntax and
the exclusion of much that was traditionally subsumed under morphology renders 'in-
flexion' an unsatisfactory translation of i'rdb, cf. Fleisch, EI, s.v. Frab.

5 Itqdn ii, esp. 260-1, 269. 6 Itqdn, naw' 36: ii, 3.
1 Volkssprache, 181.

8 Volkssprache, 165-75, see references above, pp. 86-7.

no

QURANIC STUDIES

des Nomaden ist bestimmt gesungen zu werden und der Gesang liebt
die Reibungen der Consonanten nicht, sondern braucht Vocale'. 1 Transfer
of that axiom from bedouin poetry to Muhammad's recitation of the
Qur'an (sic) may be thought gratuitous ; the source of Vrab was in any case
'bedouin' poetry. In this context it could be worth remarking that the only
cogent argument adduced by Derenbourg in favour of Vrab was that without
it Arabic poetry would no longer scan (sic): 'D'ailleurs, toutes les poesies
arabes, dont une partie a ete longtemps transmise seulement par tradi-
tion orale, perdraient le rythme qui leur est indispensable, si l'on ne re-
connaissait pas l'authenticite de la declinaison arabe'. 2

Now, whether the scansion of Arabic verse ever depended exclusively
upon regular variation of syllable length, and hence Vrab, is a question
complicated by the predominantly theoretical character of Arabic prosody:
that scansion was accentual as well as quantitative is more than likely. 3 But
Wetzstein's observations on the avoidance of consonantal friction and the
necessity of vowels are in either case, or both, apposite. In a discussion of
optional modes of poetic recitation, Sibawayh contrasted employment of
(a) tarannum, in which all final vowels were pronounced long (madd al-
sawt), with (b) the ( Tamimi) practice of adding -n (presumably tanwiti) to
all final short vowels irrespective of grammatical circumstances, and with
(c) the omission of all final vowels (save a), as in prosaic pausal position. 4
In the light of modern vernacular prosody, the attested antiquity of the
third mode is not without interest. Similarly, in a late treatise (seventh/
thirteenth century) on poetics Hazim QartajannI enumerated the devices
employed in verse to achieve aesthetic effect, among which figured the
huruf al-tarannum and/or huruf musazvwita. 5 Affixed to certain class(es) of
frequently occurring words, those vowels produced continuity of sound
(jarayan al-sawt) and transition between words (nuqla), and served also to
mark off (farq) separate themes and images. That here Vrab was meant
seems clear, not merely from context but also from employment of the
descriptive terms lawahiq and niyata (affix/suffix). Though appended
according to fixed prescription, the primary function of those vowels is not
syntactic, save possibly as thematic markers (furuq), but, rather, rhetorical,
and might be thought to correspond to that of Suyuti's non-grammatical
Vrab. 6 That distinction, however, will hardly have been other than one of
function, since Vrab vowels must have been formally identical, whether
aesthetically or syntactically applied. It may be of some interest to note
that in the examples assembled by Kofler of verse exhibiting attrition of

1 'Sprachliches aus den Zeltlagern', 193-4. 1 'Quelques remarques', 210-11.

5 See Weil, Grundriss und System, 104—5, but also 86.

4 Sibawayh, Kitab, para. 507, cited Birkeland, Pawalformen, 10-18.

5 Minhdj al-bulagha , 122-4: trans. Heinrichs, Arabische Dichttmg, 252-5.

6 See above, p. 109; and also his observation that i'rab in the sense of grammatical in-
flexion was a neologism, below, IV p. 155.

ORIGINS OF CLASSICAL ARABIC

in

those vowels, the rhetorical (acoustical) effect of i'rab is not markedly
diminished. 1

Application of the term tarannum to Qur'an recitation signified cantil-
lation, and was thus related to lahn in the sense not of defective speech, but
of melody. 2 The expression bi-luhuni 'l-'arab wa-aswdtihd employed in the
tradition adduced by Tustari and Suyuti would seem, indeed, to convey an
exhortation to employ i'rab, not because the language of Jews and Chris-
tians, and others disqualified from membership of the Muslim community,
was ungrammatical, but because it was not that of Arab bedouin. Suyuti's
use of the term lahn is, however, not unequivocal: in the passage on non-
grammatical i'rab, lahn was adduced merely as theoretical antithesis of
the grammatical variety of i'rab, while elsewhere he rejected the allegation
that lahn as lapsus calami (khata' al-kuttab) could be present in the text of
scripture. 3 This usage is naturally not that of the prophetical tradition, for
which it was proposed that luhun wa-aswat is a hendiadys meaning rhyth-
mic embellishments). Now, it may seem that discussions of Qur'an
recitation with and without i'rab could profit from attention to these several
nuances. Apart from references to the practice of Muhammad, which
complicate unnecessarily a diachronic description of CA, failure to con-
sider differentiation of the terminus technicus i'rab has meant placing
inordinate emphasis upon the evidence of both scripture and poetry for
purposes of linguistic analysis. From Suyuti's interpretation of the tra-
ditions promising celestial reward for recitation with i'rab it emerges that
such was not a matter of grammar: recitation without i'rab was, after all, to

be rewarded by half that promised for recitation with (<bj*t» jTpUl \ js ^

jive- <—ija- J5C ii vL?** °^J* ir*-) a,: - "-" v Ojj*** *^ i&

i". I; „ -^)- 4 The use of such material for serious philological argument seems
somehow unjustified. 5

Expression of syntactic relations in Arabic could not be, nor was it in
practice, limited to or even dependent upon those devices collectively and
symbolically designated i'rab. Both sequence (word order) and segmenta-
tion (conjunctive/disjunctive markers) bear a considerable portion, if not
the whole, of that burden. Incorporation of these two principles in the
styles of poetry and scripture, as well as of Kunstprosa, must and did
result in a set of modifications most conveniently described as rhetorical.

1 'Reste' (Part 3), 26-30 (case), 235-40 (mood).

1 Cf. Fiick, 'AraMya, 196-8, esp. 197 n. 16; GdQ iii, 193-4, 232-3; Hirschfeld, Re-
searches, 115.

3 Jtqdn ii, 3, and 269-75 on grammatical i'rab.
* Itqdn ii, 3 : also a prophetical hadith.

5 But cf. Kahle, "Readers', 65-71; Wehr, Review: Fuck, 181; Spitaler, Review:
Fuck, 146 ; Rabin, 'Beginnings', 25-6.

112

QURANIC STUDIES

Bloch's analysis of those styles revealed a high degree of procedural uni-
formity which ought not to be confused with what is often called 'normal
linguistic usage'. 1 Reference, explicit or implicit, to the latter for descrip-
tion of the Quranic style could seem almost insidious: if the only contem-
porary loci probantes are those found in Jdhili poetry the argument is bound
to be circular; if, however, it is conceded that the chancery papyri qualify
as contemporary comparative material, their linguistic content cannot then
fairly be judged inferior to the standard of CA. The problem can hardly be
solved by deriving that standard from the data of poetry and scripture. It
might seem superfluous to add, at this point, that in my judgement the
notions of CA origins underlying Noldeke's celebrated 'Siindenregister' of
Quranic usage are very questionable indeed. 2 Extrapolated from his
surprisingly uncritical acceptance of a pseudo-historical portrait of the
Arabian prophet, opinions about the psychological relationship of Muham-
mad to his public (e.g. 'Unsicherheit', 'Unbeholfenheit', 'Verlegenheit',
'Wiederholung als Einscharfung', 'den Arabern war eben fast alles neu',
'aller Anfang ist schwer', etc.) became organizing principles of linguistic
description. The result was a fairly systematic and thoroughly relentless
critique of a personal and individual style.

But observed from the standpoint of rhetorical schemata and organic
development, the uneven quality of scriptural style seems quite appropriate
to a document recognized by a religious community as the literary ex-
pression of divine authority. Recurrence of formulaic phraseology, for
example, might reflect conventional links between related but originally
separate traditions, rather than one man's lack of rhetorical skill. Moreover,
the recurrent formulae are functionally distinct, and their uses may thus be
ascribed to different motives, e.g. apodictic, supplicatory, narrative. 3
Similarly, repetition of rhyme words in suras 25, 4, 19, and elsewhere may
be interpreted as the same kind of linkage, extended occasionally to the
dimension of a proper refrain.* Rhyme, as well as assonance and other
paronomastic formations, operate conjunctively and disjunctively in the
manner of conventional markers: resultant constructions like those based
on epexegetic min and the juxtaposition of perfect and imperfect/past
continuous can hardly be said to cause semantic distortion. 5 In the light of
a plural anltar, the singular nahar in Q. 54: 54 could not have been pre-
ferred for the sake of rhyme, and was anyway interpreted by Zamakhsharf

1 See above, pp. 91—2, 99-100, 102-3.

2 NBSS, 5-23 ; cf. Spitaler, Review: Fuck, 146.

3 Noldeke, NBSS, 8-9; the liturgical/cultic value of qui justifies many of its occur-
rences, cf. above, I pp. 13-15.

4 NBSS, 9; see above, I pp. 13, 19, 26.

5 See below, pp. 116-17; pace Noldeke, NBSS, 9; cf. Reckendorf, Arabische Syntax,
256 para. 3a, but also 254 para. 2b for min, and 10-11 for 'prasentisch resultativ'; cf.
BSOAS xxri (1968) 612.

ORIGINS OF CLASSICAL ARABIC

113

as generic. 1 Condemnation of disjunctive formulae like p-JC*. yrjS* <U)| j etc.

as stylistic superfluities was consistent with adverse criticism in general of
recurrent locutions, but reveals in my opinion a mistaken notion of the
document's composition, and especially of its halakhic and narrative
components. 2 As for the accusation of illogic in Q. 10: 43, 27: 8o, and 30:
52, it seems abundantly clear from the context of these verses that the
epithets 'blind' and 'deaf' were intended and understood figuratively, that is
'wilfully blind and deaf to the truth'. 3 I find unconvincing Noldeke's
bewilderment at the sudden changes of interlocutor (God, angels, jinn) in
Q. 37: 164 ff. and 72: 15-16. Discontinuity of that kind is common in
Muslim scripture and may be compared with other more fundamental
lines of cleavage, such as the truncated structure of variant traditions. 4 In
Q. 72 : 15-16 the introductory particles are anyway not those employed for
the preceding speech of the jinn, and might well signal a change of speaker.

Noldeke's treatment of Quranic simile was stringent to the point of being
quite unimaginative. 5 Much of the imagery there is admittedly primitive,
but comparison of the lot of almsgivers to the increased fruits of the field
blessed by God (Q. 2: 261, 265) and that of the mean to the produce of
stony and barren soil (Q. 2 : 264) is neither illogical nor entirely unsuccess-
ful. Further, comparison of that which is spent in vanity with (the effect of)
an icy wind (Q. 3: 117) is not at all inconsistent, especially in the light of

that verse's ending: j^JJaj ^ ...ail <U)1 ^g»lk Lj. Noldeke's ob-

jection to the formulation of a secundum comparationis as instrument rather
than effect in Q. 3 : 1 17, as in 10: 24 and 18: 45, appears to me to rest upon
a very literal interpretation of metaphorical language. In Q. 2: 17-19 it is
God's extinguishing of it, not the fire itself, which must be set against the
storm producing darkness. Q. 47: 15, on the other hand, does exhibit some
syntactical confusion, owing to what must be the lengthy interpolative
gloss to janna, though it could be argued that the primum comparationis is
not janna at all, but rather the a-fa-man . . . ka-man construction of the
preceding verse. But there, and in the following example (Q. 9:19) there is,
indeed, disequilibrium between substantival constructions as first term
and finite verbal/pronominal constructions as second term of the simile.
Finally, for Q. 31: 28 I suspect that the locution ka-nafsin wahidatin is
intended to convey ka-shayin wahidin (as one thing/act). 6 It is almost
possible to understand Zamakhshari's commentary to that verse as at least

1 NBSS, 9; Kashshafiv, 442 ad loc.: ism al-jins, but it is more than likely that janndt,
here employed eschatologically, was understood as singular, see above, I pp. 27, 29.

2 NBSS, 9-10.

3 NBSS, 10; cf. Q. 2: 18-20, 7: 79, 11: 24, 47: 23; Isaiah 43: 8, Matthew 13: 10-17
alluding to Isaiah 6: 9-10.

4 NBSS, 10; and see above, I pp. 20-7. 5 NBSS, 10-11.
6 Pace Sister, 'Metaphern', 130; that schematic compilation is, however, very useful;

cf. also Buhl, 'Vergleichungen', 1-11; and below, IV pp. 239-42.

ii 4 QURANIC STUDIES

implicit support for this interpretation: JjJj jLi jLi «*Lu2 jj <jl

Inconcinnity like the parallelism of determined and undetermined sub-
stantives in Q. 42 : 49 is of course not infrequent in the Qur'an, though
this particular example appears to have been a consequence of the necessary
pausal/rhyme form, and is perhaps conspicuous only because of occurrence
in the next verse of a related locution, both members of which are un-
determined. 2 Q. 27 : 1 offers a less felicitous target for criticism (read any-
way ^bfj jTjiJI obi), since the scriptural designations qur'an and

kitab may refer to portions of or to the whole of revelation, a distinction
often expressed by absence and presence, respectively, of the definite
article. 3 To attribute some repetition to affection for thtfigura etymologica
is sound enough, 4 but Q. 4: 136 contains a fixed formula of address (L
I j^*T (jjjJI lail) which can hardly be said anywhere in the Qur'an to in-
fluence, or to be influenced by, the expression which follows. Similarly, the
particle ka-dhaiika in Q. 2: 113, 118 may be understood as presentative
'thus' and quite without influence upon the following locution (for Q. 2:

118 read ^J. 5 Q. 6: 161 5L, UJ» Up (*£^~> Jsl j*> Jl J>j JlJuk ^'1
{}S~jZu}\ ^4 oiT Uj 1^*1^1 is indeed awkward, its clumsy syntax
in my opinion a result of distinctly theological patchwork. Abraham's
role in Islamic prophetology and the doctrinal nature of the epithets
hanif and milla, rather than an aesthetic failing, might be thought to

account for this very curious construction. 6 The phrase U_i ^Ji VI

Oj-trtj hi Q. 12: 47, 48 is an appositional relative clause and may be

idiomatic. 7 Similarly, the locution «JI^». j$s which must be read in the

whole context of Q. 12: 74-5, might qualify not merely as idiomatic but
as a fair example of erlebte Rede. s Nor do I find illogical the pronominal
construction in Q. 35 : 1 1, in which it is a question of one man's life being
either lengthened or shortened. Moreover, Q. 17 : 74, which contains a
straightforward concessive construction, appears to make perfect sense
without Noldeke's paraphrastic translation. 9 Confusion of function be-
tween the concessive particles wa-in and uoa-law is a matter of degree
rather than of kind, and thus difficult to insist upon. 10 That Q. 12: 17 Lj

^jJ^Uo llf ^Jj U ^t»J*j cS\ would be logically improved by substitution

1 Kashshafui, 502 ad Q. 31 : 28. 2 NBSS, 11. 3 See above, II pp. 74-6.

+ NBSS, ir. 5 cf_ Hebrew lakhen; see above, I pp. 13-13.

6 NBSS, 11; cf. Q. 3: 66, 16: 20, and above, II p. 54

7 Pace Noldeke, NBSS, 12; cf. Reckendorf, Arabische Syntax, 435-6.

8 Pace Noideke, NBSS, 12; cf. Reckendorf, op. cit. 286.

9 NBSS, 12; cf. Reckendorf, op. cit. 495.

10 NBSS, 12, 21; cf. Reckendorf, op. cit. 494, 513.

ORIGINS OF CLASSICAL ARABIC

"5

of vca-in for wa-lato is just possible, though I am inclined, in view of the
predominantly vernacular context of Surat Yusuf, to see here the trace of
popular idiom. It may be of some interest to record that whereas Kalbl
inserted noa-in without comment, Zamakhshari resorted to paraphrase:
'even though {toa-law) we were in your opinion men of truth and relia-
bility'. 1

Two categories of Noldeke's 'stilistische Eigentumlichkeiten' are
recognized types of exegetical problem, for treatment of which specific
rules were formulated by the masoretes: change of number and change of
person. 2 The phenomena adduced here became early on part of the
established masorah, and provided the evidence from which the principle
of majazjtaqdir was derived. 3 Nowhere perhaps, does the impression of
mechanically linked prophetical logia emerge more clearly than from
observation of this discrete syntax, e.g. reference to God alternately in
first and third person. 4 The likelihood of textual mutilation as consequence
of pre-'Uthmanic methods of preservation seems to beTemote, 5 and could
hardly explain the consistent lack of logical structure in the document.
Capricious syntax would appear to reflect not the Arabian prophet's
imperfect grasp of CA, but, rather, that mechanical linkage. Noldeke's
detailed discussion of ellipsis would be even more persuasive read in the
way I have suggested, especially the 'pendant' narrative formulae. 6 Inci-
dentally, linking wa-rusulan and rusulan in Q. 4: 164-5 with inna awhayna
of the preceding verse is not quite so far-fetched as Noldeke appeared to
believe, though that lengthy enumeration probably owes its existence to a
series of arbitrary connections. 7 It may be that many instances of pleon-
asm, as of ellipsis, were the result of separate logia traditions roughly
co-ordinated. Pleonastic negation, on the other hand, often exhibits

vernacular idiom, as surely in Q. 6: 109 ^ Oc.U. lil l^Jl ^S'jxJLs Lj
OjXjJj. 8 The references to anacolutha could be infinitely expanded, 9 e.g.
Q. 7 : 2 {r±>yJS <£ j5"A3 aj jld ^j-V j ^ !As dull Jjif i_>L5"

a formulation upon which conscientious and grammatically minded
exegetes spent many words indeed. However one might interpret the
stylistic aberrations of Muslim scripture, Noldeke's conclusion that the
work was, and must remain, sui generis can hardly be challenged: 'Der

1 Tafsir, MS Ayasofya 118, 129' ad loc; Kashshaf ii, 451.

2 NBSS, 12-13, and 13-14, respectively.

3 See Wansbrough, 'Periphrastic exegesis', 254-9; a*" 1 below, IV pp. 219-27.

* See above, I p. 28. 5 NBSS, 14, 16 n. 1.

5 NBSS, 15-19. 7 NBSS, 15-16; see above, I pp. 34-5 on wahylhsdl.

8 NBSS, 19; the examples cited there, and in Zwr Gramm., 91, present fewer problems
if wa-ld is read 'or', as in vernacular Arabic; for extension of pleonastic negation see
Blau, Christian Arabic, 313-15; also BSOAS xxxi (1968) 613.

» NBSS, 22.

n6

QURANIC STUDIES

Koran bildet eine Literatur fur sich, er war ohne wirklichen Vorganger
und konnte auch keine Nachfolger haben'. 1

Whatever one could respond to the question of possible precursors, it is
very doubtful whether Quranic style ever had an effect upon the subsequent
course of Arabic literature, save as source of aphoristic citation. Its
curious system of periodization (Javidsit) made relation to the literary genre
known as rhymed prose (sap) unavoidable. The effect of that system upon
Quranic syntax did not of course escape Noldeke, and was more recently
the subject of a special study in which morphology and lexicon were also
included. z Terminological dispute about whether Qur'an and saj' could
even be mentioned in the same breath found a kind of resolution in the
expression mutamathilat al-maqaii (homoioteleuton). 3 Arguments against
the obvious were mostly of a theological or adhominem character: e.g. if the
Qur'an were nothing but traditional (sic) rhymed prose, how could it be
miraculous/inimitable (mu'jiz)} or, in saj' meaning was made to fit the
rhyme, while in the Qur'an verse endings may enhance but do not inform
the communication, etc. Resort to termini technici provided the useful
fasila, by which Quranic rhyme could be distinguished from that of poetry
(qafiya) and of saj' (qarina). In the analysis of forty kinds of grammatical
and other mutation (ahkatn) both characteristic of and affected by the
Quranic fasila, Suyuti adduced material sufficient to support the contention
that Quranic syntax, as well as certain of its morphological and lexical
features, owed something of their eccentricity to its periodization. 4 But
relation of the fasila to other forms of rhyme (i.e. qafiya and qarina) was
never suppressed: a shared descriptive terminology derived from the
vocabulary of rhetorical ornatus (bad?) was offset by the assertion that the
rules of application for the fasila were other than those which obtained for
poetry, e.g. tadmin (enjambment) and ita* (repetition of rhyme word),
condemned in verse, were recommended in prose. 5

Even a rough statistical survey of Quranic rhyme indicates predominance
of forms containing long vowel plus consonant (usually iju with -k, or iju
and a with -« or another consonant), followed in order of frequency by
vocalic rhymes (ahjajiya) and, finally, forms containing short vowel plus
consonant. 6 That the dominant rhyme pattern (long vowel/diphthong plus
consonant) should also carry stress might be inferred from the pausal
pronunciation of those syllables, whose primary characteristic is the long

1 NBSS, 22.

1 NBSS, 6, 9, io, 22; and Miiller, Reimprosa, for which see BSOAS xxxiii (1970)
389-91.

3 Suyuti, Itqdn, naw' 59: iii, 290-315, esp. 292-5; cf. Norden, Antike Kumtprosa,
909-60 on clausulalcursus; Lausberg, Handbuch, paras. 985 ff, 1052.

4 Itqdn iii, 296-301, citing Ibn al-Sa'igh. * See especially Itqdn iii, 302-15.
6 See GdQ i, 36-44; Vollers, Volkssprache, 55-80; feminine rhyme is rare; Fischer,

'Silbenstruktur', 54-5 : short vowel plus two consonants is rare; Birkeland, Pausalformen,
18-21.

ORIGINS OF CLASSICAL ARABIC

117

vowel/diphthong, rather than the accompanying consonant. Segmentation
may be said, in other words, to depend as much upon assonance as upon
rhyme. A number, though not all, of Miiller's morphological and lexical
phenomena exhibit substitution of iambic for other measures, e.g. fi"dl,
tafil, tf'dl, and especially fail. 1 The effect of such is of course greater in
contexts where the rhyme/stress/pausal forms are not too widely spaced,
less where the sound-echo is heard only at great intervals. In passages of
the latter kind, predominant in the longer suras, periodization is achieved
by insertion of fixed formulae, e.g. uoa-llahu 'azizun hakim, wa-huwa
'l-'alimu 'l-rahim, etc. Employment of formulaic systems may possibly not
account for all the phenomena adduced by Miiller, for example, such lexi-
cal items as the proper names Ilyasin and Tur Sinin. 2 On the other hand,
a document which, like the Qur'an, exhibits so much of traditional imagery
and of rhetorical schemata, deserves description in those terms, rather than
as a series of deviations from 'normal linguistic usage'. Most, if not quite
all, of the Quranic passages exhibiting post-position of fa'il forms contain
scarcely varying predications of God. 3 While post-position may be con-
trasted with pre-position in non-pausal contexts, that contrast seems to me
less important than the over-all impression of formulaic phraseology, for
which normal word order is in fact post-positional. Similarly, locutions
like huwa yuhyi wa-yumit and wa-ilayhi turja'un might well represent
crystallized formulae of cultic origin and thus not the most appropriate
evidence of irregular syntax. For the same reasons the Quranic material
hardly lends itself to assertions about misuse of tempora. 4 Employment in
scripture of the perfect 'tense' as optative or apodictic reflects ancient
tradition, however inconvenient to the formulation of normative grammar,
e.g. Genesis 17: 20 ins waim inx venom TDK vona mn.s

How much of Quranic periodization may be attributed to the use of
cultic formulae and how much to conscious adaptation of the usus loquendi
to such formulae, is probably impossible to determine. That the document
owes its undeniable stylistic homogeneity to repetition has been proposed. 6
Even rudimentary data of quantification might be thought to support this
view. Of a total 78,000 words the Qur'an contains approximately 1,850
separate lexical entries, of which about 455 are hapax legomena (nomina
propria are not included in either count). The resultant ratios, 40: 1 and
4: 1 respectively, are those found for the corresponding phenomena in
Hebrew scripture. 7 A cursory examination of any Quranic concordance
will show that it is not merely separate lexica which occur repeatedly, but
also fixed images and the rhetorical conventions associated with these. To

1 Reimprosa, 41-59, 113-28. 2 Reimprosa, 136-7. 3 Reimprosa, 112-28,

4 Reimprosa, 128-30; see Reuschel, 'Wa-kdna lldhu 'alimari, 147-53.

5 Cf. Koch, Growth, 211-12. 6 See above, I pp. 46-7.
' Cf. Ullendorff, 'Biblical Hebrew', 243.

ii8

QURANIC STUDIES

the question of frequency and distribution ratios the reply may be formu-
lated in terms of 'typical structures', or entire segments of recurring
pattern. To this end my description of the document in terms of schemata
of revelation was conceived. 1 Thematic treatment (retribution, sign, exile,
covenant) and variant traditions (prophetical mission and eschatological
promise), as well as the entire range of assimilated imagery (the vanished
nations and the battles of God, angelology and resurrection), depend for
literary expression upon a set of hardly varying phrases. The relationship
of these to the rules of normative grammar is consistently anomalous. That
relationship is exhibited also in the scriptural lexicon. Surprisingly few of
the hapax legomena are from the point of view of literary Arabic rare words,
e.g. substantives like ababil, samad, tamma, 'illiyun, etc. Most represent
familiar and practical notions, like bahath, tabassam, majdlis, jawf, sakat,
ramz, lafz, harrak, hassal, etc. One is hardly justified in assuming that
these latter did not figure in common usage during the period of the
Qur'an's composition. There is of course no reason to expect in that
document only a reflection of the common usage, or to suppose that notions
not expressed there did not exist. Those semantic sectors, and there are
many, not represented in the Qur'an may be thought indifferent, sub specie
aeternitatis. 2

But even as lingua sacra scripture must be analysed as a unit of literary
production. Guided by the subject matter itself, one has not far to seek for
the archetypes of Quranic imagery. For that the search is not less arduous.
The incorporation of Biblical concepts and imagery entailed an important
stylistic concomitant, namely, the Deutungsbediirftigkeit characteristic of
sacred language. 3 From the moment of its utterance the word of God
required exegesis. Once it had achieved canonical status, scripture pro-
duced systematic interpretation. The forms generated by that process and
the hermeneutical principles from which they were derived varied with the
needs of the community. Both constitute the subject of the following
chapter.

1 See above, I pp. 1—33.
3 See above, pp. 99-100.

2 Cf. Barr, Comparative Philology, 223-37.

IV

PRINCIPLES OF EXEGESIS

The elaboration of Islam may be seen as co-ordination of three generically
distinct factors: canon, prophet, and sacred language. A useful index to the
relative significance of each in that elaboration emerges from examination
of the kinds of exegetical literature in existence at various points of its
course. Even apart from the interpretative material included within the
text of the canon, the exegetical literature can hardly be described as
homogeneous, and several criteria have been employed in its description :
the type of scriptural material treated, argument of the author, date of
composition, etc. I propose here to experiment with two different criteria
of classification, the one stylistic, the other functional, which seem to me to
be mutually corroborative in producing the following exegetical typology:

1. Haggadic.

2. Halakhic.

3. Masoretic.

4. Rhetorical.

5. Allegorical

From the point of view of function, by which I mean the role of each in the
formulation of its history by a self-conscious religious community, these
exegetical types exhibit only a minimum of overlapping and, save for the
last-named, might almost be chronologically plotted in the above sequence.
The increasing sophistication discernible in the treatment of scripture
corresponded to a demand, at least among the exegetes themselves, for
finer and subtler terms of clarification and of dispute. Allegorical exegesis
represented a reaction to the generous manner in which that demand had
been met.

From the point of view of style, however, the types proposed here dis-
play a degree of uniformity which could be misleading. Common to all is
the presence, wholly or in part, of the canonical text of revelation. One of
the problems to be examined is the precise relationship of that text in each
type of exegesis to the accompanying commentary, qualified by the obser-
vation that redactional processes have contributed inevitably to an over-
simplified picture of that relationship. A second cause of uniformity is
posed by the recurrence in each type of technical terms whose usage seems
at first sight to be constant, thus providing apparent, but in fact deceptive,
evidence of methodological similarity. An example was application of the

120

QURANIC STUDIES

term majaz to two quite different exegetical procedures. 1 Finally, the
identity of explanations, whether lexical, grammatical, or rhetorical,
throughout the literature of scriptural interpretation provokes an under-
standable impression of uniformity. Here, too, the evidence can be decep-
tive. The value, indeed the sense, of any such explanation can, it seems to
me, be elicited only from the total context of the commentary in which it
appears. Thus, glossing alladhina kafa.ru (those who reject/disbelieve) in
Q. 25 : 32 as yahud (Jews) will not have meant to Zamakhshari (d. 538/
1 143) what it did to 'Abdallah b. 'Abbas (d. 68/687).* Similarly, the
'extrapolation' technique recommended by a number of scholars in the
attempt to recover from the works of later exegetes those of the early
authorities (e.g. Mujahid from Tabarl, al-Asamm from Tha'labi) can have
but limited success, quite apart from the presence of defective chains of
transmission and of conflicting judgements based on a single authority
(perfectly illustrated by the profusion of dicta ascribed to Ibn 'Abbas). 3
It will be useful to remember that no writer merely transmits, and
that even a compilation reveals principles both of selection and of
arrangement.

As corrective to this impression of uniformity I propose for the stylistic
analysis of exegetical literature a distinction between the elements of
explication and the framework in which they appear. Both may be described
as containing a measure of characteristic form, the relation between them
being always one of tension, often of opposition. Tension may be said to
exist when explicative elements that are minimal and to some extent basic
units of interpretation (e.g. hakadha for kadhalika, uoallahi for tallahi)
occur in a typical context, which is their original framework or point of
literary origin. Such elements, always recognizable and seldom productive,
are freely borrowed by writers employing the same exegetical framework.
Opposition may be said to exist when these and more elaborate explicative
elements figure in atypical contexts. An example is the appearance in
Farra' (d. 207/822) ad Q. 17: 1 of the story of Muhammad and the caravan
belonging to Quraysh on the road from Jerusalem, an instance of pro-
phetical thaumaturgy quite out of place in a masoretic context. 4 Less
conspicuous but no less important examples of opposition between element
and framework are found in late exegetical works where dissenting or
minority opinions were introduced by such locutions as wa-qila (and it has
been said) and wa-quri'a (and it has been read). For example, in the variae
Uctiones kalimatlkalima for Q. 7: 158 the non-canonical singular was
adduced reluctantly by both Zamakhshari and Baydawi (d. 691/1292)

1 See Wansbrough, 'Periphrastic exegesis', 354-9, 265-6; and below, pp. 227-32.

2 Kashshdfm, 278 ad loc.; Suyup, Itgdn i, 122; see above, I pp. 36-8.

3 Cf. Birkeland, The Lord guideth, 20-1, 62-3; Horst, 'Zur Uberlieferung', 290-307;
see below, pp. 139-40.

4 Ma'dni 'l-Qur'dn ii, 115-16; see above, II pp. 69-70.

PRINCIPLES OF EXEGESIS

121

and the dogmatic equivalent kaKma:hgos included only as a final
option. 1

Now, it is the ascertainable quantity of explicative elements in a typical
context or habitual framework which constitutes an exegetical type. Such a
type exhibits a structure sufficiently consistent for intrusions of the kinds
alluded to above to be noticeable, and I have indicated five which in my
judgement provide a typology of Islamic exegetical literature. The total
number of elements will of course vary according to definition, which for
the following analysis will be as broad as possible. I have selected twelve
kinds of procedural device, each of which is potentially variable but not
beyond recognition, irrespective of the combinations in which it is em-
ployed :

1. Variae lectiones.

2. Poetic loci probantes.

3. Lexical explanation.

4. Grammatical explanation.

5. Rhetorical explanation.

6. Periphrasis.

7. Analogy.

8. Abrogation.

9. Circumstances of revelation.

10. Identification.

11. Prophetical tradition.

12. Anecdote.

Study of the distribution of these phenomena across the range of exegetical
literature ought to produce a means of isolating the essentially separate
activities which preceded the appearance of classical Islamic tafsir. That
is a conventional term whose origin appears to have been rhetorical rather
than exegetical and the result of a preoccupation with profane rather
than with sacred literature. 2 It may have been the monumental work of
Tabari (d. 311/923) which contributed to an almost permanent eclipse of
some interesting controversy on the subject of exegetical nomenclature.
It is not at all impossible that the arguments of his younger contemporary
Maturfdi (d. 333/944) were calculated to establish a fixed semantic value
for the term tafsir, not merely to postulate the eventually inconclusive
distinction tafsir — ta'wil. 3 But whatever its technical designation, Quranic
exegesis would seem to have found its earliest expression within a basically
narrative framework which may conveniently be described as haggadic.

1 Cf. BSOAS xxxiii (1970) 392, on the misinterpretation of that method in Khoury,
Les Theologiens byzantins, 21 n. 13; see above, II pp. 76-7.

2 See Wansbrough, 'Qur'anic exegesis', 469-85.

3 See Gotz, 'Maturidi', esp. 31-8; and below, pp. 154-8.

122

QURANIC STUDIES

I. HAGGADIC EXEGESIS

In an exegetical work ascribed to Muqatil b. Sulayman (d. 150/767) and
entitled Tafsir, the author related the following story ad Q. 18: 9. 1

Abu Jahl said to Quraysh: Send someone to the Jews of Yathrib to ask them
whether this man (Muhammad) is a prophet or a liar — that would be a solution.
So they sent five, among whom were Nadr b. Harith and 'Uqba b. Abl Mu'ayf.
When these reached the city they said to the Jews: We have come to you about
something which happened to us recently and which is getting worse. We fear
that it may cause confusion and upset the status quo. This man— humble, poor,
and an orphan — prays to Rahman. Now the only Rahman we know is Musaylima
the false prophet who, as you also know, never caused anything but destruction
and slaughter. But he (Muhammad) claims to be informed by Gabriel, who is an
enemy of yours, so tell us whether you find any mention of him in your scriptures.
The Jews replied : We do find him described as you say. The men of Quraysh
interrupted: Among his own clan those nobler and richer than he do not believe
him. The Jews answered: We find that his own people are those most violently
opposed to him, and yet this is the time in which he is to appear. The men of
Quraysh countered: But Musaylima the false prophet is his teacher, so tell us
something that we can ask him about, that Musaylima cannot teach him, and that
only a prophet can know. So the Jews said : There are three things — if he knows
them he is a prophet, if not he is a liar. Ask him about the Men in the Cave, and
Dhu 'l-Qarnayn, and the Spirit. And they told the men of Quraysh the stories of
the first two and added, with regard to the third: If he says either little or much
about that, he is a liar. Delighted, the men of Quraysh returned to Mecca, and
Abu Jahl said to the prophet (sic): Son of 'Abd al-Mutallib, we are going to ask
you about three things — if you know them, you are telling the truth, and if not,
you are a liar. The prophet replied : What is it ? Ask me what you like. So Abu
Jahl replied: We ask you about the Men in the Cave, so tell us about them. And
Dhu 'l-Qarnayn, so tell us about him. And the Spirit, so tell us what that is.
If you know what they are, you are vindicated, but if you do not, you are deluded
and bewitched. And the prophet answered: Come back tomorrow and I will tell
you. But he did not say: God willing! He waited three days, then Gabriel came
to him, and the prophet said: Gabriel, the people are asking me about three
things. The latter replied: For that very reason I have come to you.

Somewhat abridged and following upon an elaborate description of
attempts by Quraysh to seduce Muhammad with promises of money,
power, and even medical attention, the same story appears in the biography
of the Arabian prophet composed by Ibn Ishaq (d. 151/768), in the recen-
sion of Ibn Hisham (d. 218/834). 2 In that account Quraysh sent only two
men (those actually named in MuqauTs version) ; the three minor motifs —
connection of Rahman with Musaylima, Gabriel as enemy of the Jews,

1 GAS i, 36-7; MS H. Husnu 17, i67 v -i68 r .

2 GAS i, 288-90, 297-9; Sira i, 300-2.

PRINCIPLES OF EXEGESIS

123

and prediction of Muhammad in Jewish scripture — were omitted; the
three components of the 'test' were designated Youths (fitya), Wanderer
(rajul tawwaf), and Spirit (rtth) ; and Muhammad waited fifteen days for
Gabriel.

In both works the story provided a narrative framework for lengthy
commentary on Surat al-Kahf. Examination of the means by which that
was achieved ought to throw some light on the structural consistency
of haggadic exegesis. Central to both versions is the part played by the
rabbis of Yathrib (Medina) in the formulation of its strategy by Meccan
opposition to Muhammad. As such, it is one of many but typical of all
literary devices which implied a historical link between two sources of
resistance to the Arabian prophet. 1 Employment of this particular frame-
work, which figures also as an immediate cause of revelation (sabab al-
nuzul), compels the reader to accept the Qur'an document as a source for
the life of Muhammad and thus for conditions in the Hijaz during the
seventh century. With regard to the integrity of the Quranic text,
Muqatil's use of the narrative appears at least to be the more rigorous of
the two: the entire content of Surat al-Kahf {one or two parts but tacitly)
was related to the story of Abu Jahl and the rabbis. The author achieved
that end by resort to two somewhat mechanical devices. Ad Q. 18: 1
the word 'iwajan was glossed mukhtalifan and the meaning 'variation/
irregularity' applied not to the content of scripture but to its mode of
revelation. 2 Evocation of the terminology pertinent to that dispute will have
been intentional, since Muqatil added immediately a prophetical anecdote
containing an admonition addressed to the Jews by Muhammad on the
certain consequences of their ignoring the message from God. Moreover,
the prolepsis so characteristic of Muqatil's style is exhibited here by con-
joining to the prophetical anecdote the claim of four Medinese Jews
(named) that 'Uzayr (Ezra) was the son of God. The locus classicus for dis-
cussion of that allegation is Q. 9 : 30, where it may be combined with anti-
Christian polemic, as well as with the idolatry of Quraysh. 3 Muqatil
interpreted Q. 18: 4 exclusively as reference to the Jews, while Ibn

Ishaq mentioned only Quraysh: J Uj^s l-Uj -*_Ul Jj>*i\ I^Jls ^AJI

"UJl ob a5o^VJI JUo UJ p-ftJy a formulation which, however un-
likely the utterance ascribed to the prophet's Meccan opponents, did
service to the cause of the exegetical tradition which emphasized the Hijazi
background to Islam. As such, the Sira passage could be thought to
exhibit traces of editorial intervention, while Muqatil's treatment, in

1 See above, I pp. 16-17, 18, 36, II pp. 62, 72-3, 78-80; and Hirschfeld, 'Controver-
sies', 100-16.

2 See above, I, pp. 36-8.

3 e.g. Zamakhshari, Kashshdf ii, 263 ad loc.

124

QURANIC STUDIES

contrast, is almost certainly a reflex of Rabbinic and/or apocryphal tradi-
tions concerning Ezra. 1

Thus, in Muqatil's version verses 4-8 of Sura 18 were anticipated and
the narrative, replete with allusion to the fate of earlier heedless peoples
(umam khaliya), 2 brought up to the account of Abu Jahl. Two of the three
episodes representing the components of the 'rabbinical' test of prophet-
hood, contained in verses 9-26 and 83-98, are commented at some length,
and the third accommodated by reference at verses 108-10 to yet another

facet of the Jewish opposition to Muhammad: 3 ^joyj^i ^i-^IU.

tii*| 6jS"i tJ^' <J& dy^i 1-*^ Jjj j~>\ iy

C.U15" JLtUi' jf J-J ^aaJI AiJ oU153. This passage (obliques
separate canonical text from commentary) illustrates several aspects of
Muqatil's method: zero connective between hhalidin fihd and layamutun
may signal paraphrase; the connective ya'ni between hiwalan and muta-
hawwilan may introduce either an interpretation or, as here, a gloss;
wa-dhalika an/anna generally indicates the 'occasion' of revelation;
fa-qdla subhanahu signals resumption of the canonical text, often accom-
panied, as here, by a 'stage direction' indicating the person(s) addressed.
Now, in this instance Jews were adduced expressly to complete the narra-
tive: by including the third component of the test of prophethood, namely,
knowledge of the Spirit. But rather than cite the relevant Quranic
passage (17: 85), Muqatil paraphrased it in terms of a Jewish allegation,
Muhammad's assertion, and God's confirmation of His prophet, in the
course of which scriptural syntax suffered not a little. The stylistic advan-
tages of this procedure become apparent when compared with the method
of Mutatil's contemporary, Ibn Ishaq.

There, a longer and more intricate prelude to the revelation of Surat
al-Kahf necessitated a correspondingly complex interpretation, in the
course of which fragments of ten Quranic suras were treated (in the follow-
ing sequence: 19: 64, 18: 1-8, 18: 9-26, 18: 83-98, 17: 85, 31: 27, 13:
31, 25: 7-10, 20, 17: 90-3, 13: 30, 96: 9-19, 34: 47, 41: 26, 74: 31,
17: 110), being adduced as the revelations brought to Muhammad in
consequence of his encounters with the combined forces of Arabian Jewry

1 Tafsir, MS H. Hiisnii 17, 167", Sira i, 302; see below, p. 127; Geiger, Was hat
Mohammed, 191-2; Horovitz, Untersuchungen, 127-8, 167; Speyer, Erzahlungen, 413;
and cf. Kiinstlinger's speculation on the messianic content of 4 Ezra 7 : 28-9, "Uzair",
381-3, though attribution to Muhammad of a 'mistaken interpretation' seems far-fetched.

2 See above, I pp. 2-5.

J Tafsir, MS H. Hiisnii 17, i67 v -9 r , i72 r -3 r » i73 M , respectively.

PRINCIPLES OF EXEGESIS

125

and Quraysh. 1 Ibn Ishaq's treatment is characterized by the strictest
economy. The three episodes central to the theme of the Medinese rabbis
occupy six pages, and the scriptural passages (18: 9-26, 18: 83-98, 17: 85)
are compactly and explicidy presented. 2 Narrative precedes and follows
but does not interrupt. Interpretation consists primarily of gloss and
paraphrase, the only connectives employed being ay and zero. Ad Q. 17:
85 Jews and Quraysh were brought together in the following manner:

ajj.u!I <UJl Jj^j fJS U Jl» aJI (.rUt ^1 <Jio.uj (jjWJ <JU

jj^J UU ^1 jJuJI ^ pZJjl Uj/dLly c-jfjf JU>^ L jU.f cJls

JS" ilj^zJI LJjt jj Ul i3 ft U UJ jk: db'ls IjJls <Jtf «iLy

jj j^XjiJo L dAJj J ^TjUcj JJls «UJ| JLc L^Jt <Ujl Jj^.j Jlai

ejA^Jii. There the rabbis (ahbar yahud) confronted Muhammad in

Medina with an alleged utterance of his confirming the omniscience of
the Torah (formulated as revelation: tatlu fima jaaka) and citing
(correctly) Q. 17: 85, to which Muhammad retorted that it referred to both
Quraysh and Jews: and of knowledge you have been granted but little.
Thus the story of Abu Jahl and the rabbis was completed and confirmed by
a later event. In Ibn Ishaq, moreover, the denouement was logical and
consistent. In Muqatil the Jews reproached Muhammad for claiming to
have knowledge from his Lord and yet professing to know nothing of the
Spirit, thus revealing their treachery, since it was that very ignorance
which was to be proof of Muhammad's divine calling. 3

The 'rabbinical' test of prophethood employed with varying skill by
the exegetes is clearly remote from Jewish doctrine, and reflects a very
primitive level of polemical discourse. 4 But its function here might be
thought stylistic rather than merely polemical, a suspicion corroborated
if not confirmed by the transparent adaptation of an ayydm motif: the
offer of three courses of action (thalath khisdl) to the protagonist, as a
means of both stimulating and limiting movement within the narrative
framework. s The motif was consistently related to Jewish, if not always
to Meccan, resistance to Muhammad, as can be seen from a version adduced

by Suyuti: cH)jL, ^ JUs eU'ti <U)| <Jj-«j ^}*u« ^ Xs- JU

AiaJI J&f pUL Jjt Lj itLJI <Jjf L "ill <£Otf ,jP

J15 Jj^. JlS UiT Jj^o. ^ ^ ti' -?f *J J 1 {jri W

1 Sira i, 294-302 and 302-14. 2 Sira i, 303-8.

3 Cf. Geiger, Was hat Mohammed, 38 ad Q. 25: 4-

* See above, II pp. 70-1, 74; and cf. Vajda, 'Juifs et Musulmans', 99-108 for the
versions in Ahmad b. Hanbal.

5 See Caskel, 'Aijam al-'Arab", 49-52; Widengren, 'Oral tradition', 232-43; Sellheim,
'Die Muhammad-Biographie', 70, 84-5; Stetter, Topoiund Schemata, 36-9; cf. the same
device in 1 Chronicles 21 : 10-13.

126

QURANIC STUDIES

(Q. 2: 97) 3L$I tj3 a5C^JI & ij^JI jo* dJi Jli There the
interrogation took place in Medina and the three components of the test
were different, but the minor motif depicting Gabriel as enemy of the
Jews was included, with the expected reference to Q. 2: 97. 1 Elsewhere
Suyuti adduced a part of the Muqatil/Ibn Ishaq version of the test, in-
cluding only Q. 17: 85 and the question of the Spirit, with a choice of
Medinese or Meccan setting, and with characteristic logic opted for the
latter since that was where Q. 17 : 85 had been revealed. 2

The observation of Noldeke-Schwally: 'das Ganze ist hochst fabel-
haft, so dass wir nicht viel darauf zu geben haben', is also characteristic of
the arbitrary 'historical' method which for a century has dominated
the course of Islamic and particularly of Quranic studies. There are, in
fact, very few verses of the text of revelation which were not 'bald nach
Mekka, bald nach Medina verlegt', abundantly clear from even a cursory
reading of the first-fifteen chapters of Suyutl's Itqdn. 3 Despite Noldeke's
confident assertion: 'Wir haben vor ihnen alien aber doch namentlich eins
voraus: die Unbefangenheit gegeniiber dem religiosen Vorurteil. Und
dazu sind wir in der Schule der wissenschaftlichen Kritik aufgewachsen',
his historical evaluation of traditional data did not bring him much
beyond the position established and occupied by Suyuti 400 years earlier. 4
Modifications of Noldeke-Schwally by Bell and Blachere, respectively,
exhibit refinement of detail but no critical assessment of the principle
involved, namely, whether a chronology/topography of revelation is even
feasible. 5 Nor is the historical analysis of six suras undertaken by Birkeland
free of the implications of that principle, which can, after all, only be
a matter of conjecture. 6 An example is his interpretation of Q. 93 : 6-8
in terms of exegetical alterations inflicted upon the data of the prophet's
'orthodox' biography. 7 A literary analysis would at least require con-
sideration of traditional cultic formulae, e.g. Psalms 10 : 12-18, 22 : 24 on the
'orphan's lot'. 8 Suyutl's argument for Q. 17: 85 was anyway of minimal

1 Itqdn i, 97: citing Bukhari on the authority of Anas; see above, II pp. 62-3.

2 Itqdn i, pp. 93-4.

3 GdQ i, 139: but both rejection and acceptance of these conflicting reports presuppose
criteria of assessment at worst pernicious, at best subjective, see above, I pp. 38-41, and
below, pp. 177-81. * NBSS, 5-6.

5 Cf. Bell, Bell's Introduction, esp. 108-20; Watt, 'Richard Bell's theories', 46-56;
Blachere, Introduction, esp. 182-98, 240-63 ; but also Torrey, Foundation, 91-8.

6 The Lord Guideth, suras 93, 94, 108, 105, 106; id. 'Surah 107', 13-29.

7 The Lord Guideth, 23-37.

* Such is the method attempted here, particularly in ch. I (theodicy) and ch. II (pro-
phetology). A valuable and detailed exposition of the snares inherent in literary analysis
may be studied in the recent work of Richter, Exegese als Literaturwissenschaft, esp. 27-48.
For the Quranic revelation a systematic theology, such as Stieglecker, Die Glaubenslehren
des Islam, or a comparative one, such as Masson, Le Coran et la revelation judeo-chretienne,
dispenses with the historical dimension and thus also with a makeshift psychoanalysis of
prophetical experience, see above, II pp. 56-8.

PRINCIPLES OF EXEGESIS

127

significance for the narrative value of the prophetical 'test' ; concern with
the precise dating and location of separate revelations was a serious occupa-
tion only at the halakhic level of scriptural exegesis.

For Muqatil and Ibn Ishaq it was the story that mattered. Indeed, it
may be said of the former that the scriptural text was subordinate, con-
ceptually and syntactically, to the narratio. That this is less true of Ibn
Ishaq's work could be a result of its having been drastically edited by a
scholar fully conversant with the methods and- principles of masoretic
exegesis. To anticipate with a single illustration my discussion of that
exegetical type: in the Sira lexical problems, e.g. bakhi'un nafsaka (Q.
18: 6), al-raqim (18: 9), shatafan (18: 14), are elucidated by reference to
loci probantes from poetry, not, however, by the author of the work but
by its editor (signalled qala Ibn Hishdm). 1 When glossing a scriptural
locution, e.g. sultan bayyin (Q. 18: 15), Ibn Ishaq merely declared: that
is, an eloquent proof {ay bi-hujja bdligha), without adducing external
evidence. 2 Muqatil, on the other hand, limited his comparative material
for lexica to scriptural shawahid, introduced by the expressions wa-
naziruha (analogous to that) and mithla qazolihi ta'dld (as in scripture),
which became technical terms in the masorah. 3 Now, the intrusion
of such editorial elements into Ibn Ishaq's biography of the prophet is
only the beginning of an answer to the question: how did the author of the
Sira deal with scriptural material ? From my analysis of the story of Ja'far b.
Abl Talib and the Najashi it will be clear that I regard the narratio as
paramount and the isolation of scriptural texts in canonical form an after-
thought. 4 The fairly tidy separation of scripture from narratio in the
story of Swat al-Kahf I am tempted to ascribe to editorial revision of the
kind in which poetic shawahid for scriptural lexica would also be charac-
teristic. In other words, the Sira exhibits evidence of both halakhic and
masoretic reformulation: of the former in its attention to asbdb al-nuzul,
nowhere more clearly illustrated than in the proliferation of Quranic
passages appended to the story of Abu Jahl and the rabbis of Medina;
and of the latter in the employment of poetry to explain Quranic lexica.
None the less, the structural similarity between the works of Ibn Ishaq
and Muqatil seems to me almost beyond dispute, and if the terms sira
and tafsir later became designations of distinct literary genres, their basic
identity for the earlier period may, I think, be conceded. 5

The narrative style is best observed in Muqatil. There, separation of
scripture from commentary is difficult, frequently impossible. The author
achieved that unity of presentation by resort to several devices : prolepsis,
repetition (subsumption), presence/absence of connectives, stage directions,

1 Sira i, 302-4; see below, pp. 216-18.

3 See below, pp. 208-15.

5 Cf. BSOAS xxxi (1968) 148-9.

1 Sira i, 304.
♦ See above, I pp. 38-43.

128

QURANIC STUDIES

supercommentary, and interpolation (paraphrase). Prolepsis is the means by
which the narratio was maintained intact, and I have described the manner
in which the first eight verses of Siirat al-Kahf were rendered introductory
to the tale of Abu Jahl, with subsequent accommodation of verses 9-26
(the Men in the Cave), verses 83-98 (Dhu '1-Qarnayn), and inclusion of the
third component of the test (the Spirit) by reference at verse 109 to a con-
frontation of Muhammad with the Jews of Medina, in which Q. 17 : 85
was paraphrased (equation of kalimat rabbi in 18: 109 with 'Urn in 17:
85). Intervening segments of the sura were related to the narration (Leit-
motiv) by a system of cross-reference. For example, the parable of the
two gardens, verses 32-44, was connected with the Men in the Cave by
giving the name of one of them, Yamlikha (who was sent by the others
into a near-by village to buy food and thus brought about their discovery),
to that one of the (two) gardeners who found favour with God. Now, the
(two) gardeners were members of B. Isra'il and so too, according to Muqa-
til, were the Musa and Khidr of the journey related in verses 60-82.
Identification of the latter was important, since it appears to have been
a matter of dispute and intimately linked to the role of Moses in Muslim
prophetology. 1 It may also be noted that in the anecdote introducing the
journey that particular Moses was instructed by Gabriel, despite his
hostility to the Jews ( !), on how to reach the Fountain of Life ('ayn al-
hayat). 2 Of some interest, too, is the dialogue preceding the initial quest
of Musa for Khidr, in which the former was reproached by God (a non-
Quranic revelation introduced by fa-awha 'llahu ilayhi) for thinking
himself the most intelligent of men. That theme was resumed in the
ensuing conversation between Musa and Khidr, in which the former's
knowledge of God was compared with the amount of food which a bird
could with its beak collect from the sea. 3 The specifically marine imagery
of that passage ('ilm .bahr) anticipated the phraseology of verse 109 which,
as I have shown, was Muqatil's method of introducing Muhammad's
knowledge of the Spirit into his commentary on Siirat al-Kahf.

Continuity was also achieved by frequent repetition of scriptural passages
occasionally paraphrased and/or anticipated by paraphrase, as in the anec-
dote introducing the journey of Musa and Khidr, where the locution
f a-awa ila 'l-sakhra anticipates verse 63. Smooth transition from com-
mentary to text could result in omission, e.g. of the anaphoric fantalaqd
hatta idha of verse 71 (cf. verses 74, 77), sacrificed in the interests of a

1 Tafsir, MS H. Hiisntt 17, i69 y -7o r , i7i r -2 r ; see above, II pp. 56-7, 76; and
Goldziher, Studien ii, 163: citing Bukhari, §ahih in, Kitab al-Tafsir, 277-82.

2 Cf. above, II pp. 62-3; Geiger, Was hat Mohammed, 168, 187-8; Zunz, Vorirage,
137-9, esp. 138 n. (a) : for Elijah and R. Joshua b. Levi; but also Horovitz, Untersuckungen,
141-3; GdQ i, 141-2; Speyer, Erzahlungen, 238-9; Schwarzbaum, 'Theodicy legends',
119-69.

3 Tafsir, MS H. Husnu 17, 171'-".

PRINCIPLES OF EXEGESIS

129

1 racy dialogue between Khidr and the ship's captain; or at verse 47, where
the gloss fa-lam yabqa minhum ahad appears without the corresponding
text. 1 Inconsistent employment of connectives (mostly ya'ni, occasionally
yaqiil, rarely ay) produces the impression that such were used not so
much or at least not exclusively to separate text from commentary, but
rather as punctuation, the equivalent of pause in oral delivery. This
usage is especially remarkable where ya'ni signals a gloss not of a scriptural
term but of one in the commentary, e.g. wa-lakinna 'uzayr 'abd allah
dakhir ya'ni saghiran, anticipating verse 4, or jazaan kariman ya'ni 'l-janna,
glossing ajran hasanan of verse 2? Supercommentary of that kind, for
example, adverse 54: sarrafna ya'ni lawwanna ya'ni wasafna, and passim,
would seem to indicate oral delivery. 3 Zero connective, on the other hand,
e.g. ad verses 108-9: khdlidin fihdjld yamiitun, and verse 31: asdwir min
dhahabjwa-asdwir min lu'lu', provokes a different kind of problem, namely;
whether the explicative element is to be understood as gloss or as inter-
polation or, indeed, as part of scripture. 4 The phenomenon is not limited
to Muqatil: in Ibn Ishaq ad Q. 18: 2 the intrusive phrase wa-'adhdban
aliman fi 'l-dkhira is glossed as though it were scripture. 5 Even more
arresting are revelations not now part of the canonical text introduced by
formulae such as wa-qdla subhdnahu and the like, usually reserved by
Muqatil to signal resumption of the canonical text in combination with a
'stage direction', e.g. qdla likuffdr tnakka, lil-yahud, til-nasara. For in-
stance, the passage ad verse 45 U^j c-~JI J^Ui^UjJI JJU ■GU^ J_^aj
Sy.St Os.U lil L'jJI diljsi ^^.j Jiyb j| y> could be

a gloss, possibly a paraphrase of Q. 6: 99, though I am inclined to under-
stand it as an independent utterance, 6 like that included in a prophetical

hadith at the end of Surat al-Kakf: 7 ^ dyy: lit «U)1 J>I>

UJU Jj^L VIJJI Vj Sijr^ *K J** 31 J*-* J- Subsumed

by traditional scholarship under the heading hadith qudsi, dicta of this sort
may owe their origin to haggadic exegesis."'

Muqatil's style is characterized by recurrence of certain minimal units
of explication whose distribution was naturally determined by the text of
scripture, but which may be regarded as constants in the over-all structure.
Examples are dhdlika glossed hddha, kadhalika.hakadha, ladun.'inda,
lawld:halld, khayr:afdal, mathahshabah, etc. Such units, in so far as they
were seldom if ever points of departure for extended interpretation, were
non-productive and applied with considerable freedom throughout works
belonging to the haggadic type. At the levels of halakhic and masoretic

1 Tafsir, i7i r , I7i v , i"jo T , respectively; the third example may be merely a lapsus calami.

2 Tafsir, 167'. 3 Tafsir, iyo y . * Tafsir, i73 r , i6g v .
5 Sira i, 302. 6 Tafsir, 170 1 . 1 Tafsir, 173".
8 See Tahanawi, htilahdt, 280-1 ; GdQ i, 256-8.

43S9C75 F

130

QURANIC STUDIES

exegesis these minimal units might become productive and even crucial
to a solution of a juridical or textual problem, but within the haggadic
framework their cumulative effect cannot be said to disturb the central
position of the narratio. Thus, examination of the exegetical work of
Muqatil's contemporary, Muhammad Kalbl (d. 146/763), 1 turns up a
similar range of minimal units of explication, e.g. dhalika glossed hadha,
kadhalika : hdkadhd, lada : 'inda, in : ma, Id alia : likay, wa-in : wa-qad, atd : a' id,
tattdhi-.walldhi, khayr-.afdal, etc. In neither writer are these syntactical
and lexical equivalents accompanied by loci probantes, scriptural or pro-
fane, or any other argument of justification.

This manner of glossing generated in the works of both writers two
characteristic and related techniques: serial repetition and circular expli-
cation. An example of the first is found in Muqatil ad Q. 18: 85, 89, 92,
where the refrain fa-atba' sababan (so he pursued a course) is glossed at
each appearance: ya'ni 'ilm asbdb mandzili 'l-ard wa-turuqihd (that is,
knowledge of the structure of the world and its ways). 2 In Kalbl ad Q. 12:
25-8 the verb qadda (cut, tear) is glossed four times by the synonymous
shaqqa, and the locution min dubur (from behind) twice by the synonym
min kkalf. 1 Again, ad Q. 12: 28 kayd (wile) is glossed twice makr wa-
sani' (deceit and deed), and immediately thereafter, ad verses 33-4 by
makr (deceit) alone; earlier, in verse 5, kayd is glossed by another syno-
nym: hila. 4 Ad Q. 12: 85 qdlu (they said) is provided with the 'stage
direction' wulduhu wa-wuld wuldihi (his children and grandchildren),
repeated for verses 95 and 97 and varied for verse 96 with the synonymous
li-banihi wa-bant banihi. 5 Now, such repetition without appreciable interval
seems quite unnecessary, and must not be confused with genuinely
helpful explanation like the gloss of aktharu 'l-nds (most of the people)
everywhere in Surat Yusuf (12: 21, 38, 40, 68) as ahl misr (Egyptians),
except at verse 103, where the admonition is transferred from a typological
to an aetiological plane and the locution glossed ahl makka (Meccans). 6
Moreover, in the last example the intervals between occurrence of the
phrase in scripture might seem to justify repetition.

The second glossing technique shared by the two exegetes, and which I
have called circular explication, may be seen in Muqatil ad Q. 18: 104-5.
There dalla (errant) is glossed habitat (in vain), while in the very next
verse habitat in the text of scripture is glossed with the synonym batalat. 7
In Kalbi ad Q. 12: 31, 51 hdsha lillah (God forbid!) is glossed each time
with the synonymous ma'ddh allah, while in verses 23 and 79, where the

1 GASi, 34-5; MS Ayasofya 118.

2 Tafsir, MS H. Hiisnii 17, 172"; at verse 8s the scriptural text itself is omitted, probably
owing to homoioteleuton with the preceding verse, but the gloss is there.

3 Tafsir, MS Ayasofya 118, i30 r . t Tafsir, i3o v , I28 T .
5 Tafsir, 134", I35 r - V . 6 Tafsir, i2o v , I3i r - V , I73 v , 135*.
' Tafsir, MS H. Hiisnii 17, 173-".

PRINCIPLES OF EXEGESIS

131

latter expression is scriptural, it is glossed both times a'udhu billah (I take
refuge with God). 1 This circular, or perhaps more accurately, sliding
explication was applied also to those glosses which I have designated
minimal units, e.g. la' alia (perhaps) is throughout Surat Yusuf explained
by Kalbl as equivalent to likay (in order to), but in verse 83 the word
'asa (perhaps) is gratuitously glossed la'alla. 2 The semantic principle that
words have uses, not merely meanings, is clearly one with which both
Muqatil and Kalbl were familiar. On the other hand, the exaggerated
manner in which those simplistic lexical equivalents were adduced pro-
vokes a question at least as to their purpose. Here I suspect that the
answer can be more profitably sought in the authors' concern with ease of
delivery, less in their concern to elucidate scripture. 3

Like Muqatil's, Kalbi's work has traditionally been entitled Tafsir,
though it is unlikely that either author called his work by that name. In
addition to the stylistic devices common to both and already mentioned,
there are other similarities but also differences. To illustrate Kalbi's
method I have selected his presentation of Surat Yusuf, an effort at sus-
tained narrative commentary which could thus dispense with imposition
of the haggadic framework (narratio) noted in Muqatil's treatment of
Surat al-Kahf. But Quranic narrative is nothing if not elliptic, often
unintelligible without exegetical complement. This structure applies
without reservation to the Quranic story of Joseph, 'the most beautiful
of tales revealed in that book' (cf. Q. 12: 3). An idiosyncratic and very
conspicuous feature of Kalbi's style may be seen in the distribution of
commentary in relation to scriptural text. Words, occasionally phrases,
even clauses, but never sentences or entire verses, were glossed in sequence
at a ratio of approximately 1:1, resulting in a highly segmented composi-
tion. This is accentuated by employment of zero connective with 'envelop-
ment' of the (preceding) text, so that separation of text from commentary
by means of obliques produces the following, very typical, pattern for

q. 12: 50-7. 4 Ij^x* o^jVu^j <i /<-*».>* ij^i/iJu.jJ duifj
^jui/yjdi vly o^W^i viy/l^i J-iHj o^ J]

L^\yi\j iiyUlj jiSOl/jjEj Ijj^j/Jrplj <U!U/lyJ". Curi-

ously, segmentation did not necessarily entail fragmentation, and it
may be that the overlapping effected here by recourse to zero connective
and retroflexive envelopment was devised by the author to ensure con-
tinuity. Though zero connective is of highest frequency by far in Kalbl, he

1 Tafsir, MS Ayasofya 118, I30 v , 132*, 130', I34 r - 2 Tafsir, 134".

5 See below, pp. 144-8. * Tafsir, MS Ayasofya 118, i32 v -i33 r .

132

QURANIC STUDIES

employed occasionally and indifferently yaqul, as well as its variation ya'ni,
the latter especially and for obvious reasons following qdlajqdlu (he/they
said), a convenient device in a sura consisting almost entirely of dialogue,
e.g. ad verse 44: qdlu ya'ni 'l-'arrdfin wal-kahana wal-sahara (they said,
i.e. the seers, the soothsayers, and the sorcerers), or verse 74: qdlu ya'ni
fata yiisuf (they said, i.e. Joseph's men). 1 The connective ay may also
occur, not as a means of joining comment to scriptures, but rather, of

adding supercommentary, e.g. ad verse 6: 2 /diuj do-^/l-ASGb/dWiT^

noted that the imagery of the two passages cited here, with particular em-
phasis upon prophethood (nubuzvwa) and its soteriological fulfilment
(islam), is identical, and thus anticipated Kalbi's concluding observations
on this sura, concerned specifically with Muhammad and the Quranic
revelation, 3 very similar to Muqatil's paraenesis at the end of Surat al-
Kahf*

Despite the essentially narrative structure of Surat Yiisuf, Kalbi's
treatment cannot be called anecdotal in quite the sense that the epithet
may be applied to Muqatil's work. There are at least two reasons for the
difference of quality between the two styles. The first is the presence in
Kalbi of two features virtually absent in his contemporary, namely,
variae lectiones and alternative glosses for a single locution. Both figure in
such quantity with Kalbi as to be characteristic of his work as preserved
and, like the presence in the Sira of asbdb al-nuzul and poetic shawahid,
may be regarded as evidence of editorial reformulation. Features such as
these, though not typical of haggadic exegesis, may exist there even in
quantity without altering the typological structure. All four devices,
together with a few others yet to be mentioned, are appropriate to hala-
khic and masoretic exegesis and thus, within the haggadic framework,
represent intrusions. That intrusive quality is clear not only from the
breach in the narrative caused by the presence of, say, a varia lectio,
but also from the nature of the element itself. To anticipate again my
discussion of masoretic exegesis with a single example from Kalbi's

work: at Q. 12: 63 the author observed /jXj/uft-^ OiA»Lif ^ J~jt»

useful illustration

of method, since it includes a haggadic element (our brother: Benjamin),
a gloss depending upon an implied textual variant (he obtains measure/

1 Tafsir, i32 r , i34 r . 2 Tafsir, 128* a Tqfsir, 136'"* ad verses 109-11.

♦ Tafsir, MS H. Husnii 17, I73 r ~ v .

PRINCIPLES OF EXEGESIS

133

we obtain measure), and an alternative (introduced 'and it is said') based
on the other of the two transmitted readings. 1 But the passage is elliptic
and the key to the argument missing, namely the reading yaktal (he
obtains/will obtain measure). For the discussion in complete and rather

more logical form, one may turn to Farra': 2 J^Cl U U.I J.*yti «Uy

JlS ,y* V*! [J^j a-UI -U& (_jt>tys! ]j>

A^U. aJ Jjiill Jjcji A*Ju) 4~-y3J J3o Jls ^jwj (JJ3I (i aLo.

j-oo J^i" Aj dj^ji There the source of both readings is stated to-

gether with the judgement of Farra' that both were correct ( !), depending
upon whether the amount of grain obtained by Jacob's sons in Egypt
was merely that promised by Joseph if they were accompanied by
Benjamin, or to be increased by a special allocation to the youngest
brother. Now, almost without exception the textual variants adduced by
Kalbi are of elliptic and referential character, and appear to me to pre-
-suppose acquaintance with the masoretic activities of scholars like Farra'.
The implications of this hypothesis, which touch upon matters of red-
action, will be examined in due course. 3 It may suffice here to suggest,
despite absence of explicit editorial revision of the kind available for the
Sira, that Kalbi's work as preserved exhibits a considerably modified form
of haggadic exegesis.

The second cause of difference between the commentaries of Kalbi and
Muqatil is the absence in the former of a narrative device much favoured
by Muqatil: the prophetical tradition (hadith). In a work which, like
Kalbi's, was claimed to have been transmitted exclusively from the
authority of 'Abdallah b. 'Abbas, one might expect a profusion of such
dicta, as well as of hadith qudsi, also remarked in Muqatil's work. In the
latter the prophetical hadith could be didactic, as for Q. 18: 46:*

And the Prophet said: The enduring good works are [utterance of the prayers]
'Glory to God', 'Praise be to God', 'There is no god but God', and 'God is great'.

Or it might be merely anecdotal, an entertaining digression, as for Q. 18:
96:5

And a man said to the Prophet: I have seen the barrier of Yajuj and Majuj. So
the Prophet said : Describe it to me then, And he replied : It is like striped cloth —
black and red. The Prophet said: Yes, you have seen it.

The absence of this kind of material in Kalbi is difficult to explain. In
contrast to the unhurried, almost chatty style of Muqatil, Kalbi is terse,
humourless, matter-of-fact. That austere manner may have been inten-
tional, but comparison of the two authors together with the significant

« Tqfsir, MS Ayasofya n8, i35 r . 2 Ma'dnl 'l-Qur'dn ii, 49 ad loc.

» See below, pp. 138-44. 4 Tqfsir, MS H. Husnu 17, 170'.

s Tafsir, 172*.

134

QURANIC STUDIES

fact that both preceded by nearly two generations that stage of exegetical
scholarship in which the basic problems, doctrinal and textual, were
to be examined in exhaustive detail, provokes a question at least about
the redaction of Kalbl's commentary.

The narratio was not, however, entirely obscured. The Quranic story
of Joseph requires, to escape the stigma of nan sequitur, a minimum of
supplementary material. Such was provided by Kalbi, and in a proleptic
form reminiscent of MuqatiPs method. For example, ad Q. 12: 36 the
dreams of Pharaoh's cupbearer and baker, only alluded to in the text of
scipture, were set out in considerable detail and also interpreted by
Joseph, thus anticipating verse 41. Moreover, the interpretation contains
several refinements not found in Genesis 40: 9-19, e.g. the symbolic
values of the vineyard and of the vine, as well as that of the three branches. 1
Again, at 12: 43 Pharaoh's dream was embellished to accord with Joseph's
interpretation of it at verse 49, and twice provided with the interpolation
'emerging from the river' (^gj ^ cf. Genesis 41: 2 nbs 1X71 p). 2

At 12: 59 Joseph's peremptory demand: Bring me a brother of yours from
your father, was supplemented by the (very necessary) 'as you have just
said that you have a brother at home with your father', an interpolation
expanded without comment at verses 69-70: '(Joseph's) brother, from
the same father and mother'. 3 At 12: 93: Take this shirt of mine,
was glossed 'and his shirt was of heavenly origin', a reflex of Rabbinic
descriptions of the CDS niro (Genesis 37: 3). 4

The obvious source for most, if not all, of that material is Rabbinic
literature, which has been culled and collated by a number of scholars,
among whom the most knowledgeable and sophisticated were Horovitz
and Speyer. 5 For the document of revelation itself, the latter 's work might
be described as exhaustive, but an examination of Muslim exegetical
literature would have revealed additional and equally interesting parallels,
and prevented at the same time one or two false impressions. To declare,
for example, of the locution qdla kabiruhum in Q. 12: 80 that 'Es sprach der
grosste von ihnen, wobei mit kabir hier sicherlich der alteste, also Ruben,
gemeint ist' may be logical (cf. Genesis 42: 22) but is quite unnecessary:

1 Tafsir, MS Ayasofya n8, i3i r . 1 Tafsir, i3i T -2 r . 3 Tafsir, i3?~ v .

4 Tafsir, i3S r ; and cf. Speyer, Erzahlungen, 219-20.

5 e.g. Geiger, Was hat Mohammed; Griinbaum, Sagenkunde; Schapiro, ElemaUe;
Sidersky, Ligendes; Katsch, Judaism; cf. Heller, 'Recits', 113-36 (on Basset); id.
'Legende', 1-18 (on Ahrens, Torrey, and Kunstlinger, but especially Sidersky); for the
problematic assumption of 'emprunts' see Moubarac, Abraham, 163-75 ! m y concern here
is less with Biblical and Rabbinic elements in the Quranic text than with the materials
and methods employed in the elaboration of Quranic exegesis: haggadic material from
later exegetical works (post-Tabari) was adduced by several of these scholars, often
working independently, of which the widest selection may be found in Griinbaum,
Sagenkunde (e.g. Ya'qubi, Tabari, Mas udi, Zamakhshari, Baydawi, Ibn Athir, Abu
'1-Fida, Qazwlnl).

PRINCIPLES OF EXEGESIS

135

^ Kalbi, as well as most of his successors, knew that it was Judah, 'the fore-
most of them in intelligence' (cf. Genesis 44: 18-34). 1

Thus the gaps in the Quranic narrative were filled from a very familiar
mine of Biblical lore. But only at Q. 12: 36 and 43, those verses having to
do with interpretation of dreams, does Kalbi's commentary exceed in
length the portion of scripture being interpreted. Though it could be
argued that the laconic and somewhat monotonous style is witness to the
author's conviction that his public required no further amplification of
the well-known story, such is not really supported by the available evidence.
I have mentioned the identification of Judah at Q. 12: 80. He it was also,
who at 12: 10 counselled his brothers not to kill but to sell Joseph, and
who at 12: 96 brought the news of Joseph's existence in Egypt to Jacob. 2
At 12: 8 and 63-76 the brother was of course Benjamin; at 12: 4 Joseph's
parents were identified as Rachel and Jacob; at 12: 19 the scene of the
crime was Dtishan (Dothan); at 12: 21 Joseph's buyer in Egypt and his
wife were named Quzifar (Potiphar) and Zulaykha; at 12: 36 the two
'servants of the king' imprisoned with Joseph were specified vintner and
chef, later cupbearer and baker; and at 12: 93 the number of Joseph's
family sent for from Palestine was put at about 'seventy persons'. 3 Now,
save for Zulaykha and placing Dothan between Midian and Egypt, this
material is unexceptionable. There was some further embellishment: at
12: 19 the man who rescued Joseph from the well and sold him in Egypt
was called Malik b. Daghr, an Arab (bedouin) from Midian; at 12: 30
and 50 the women responsible for Joseph's humiliation and imprisonment
(four in number) were identified by their husbands' respective ranks in
Pharaoh's service; and at 12: 94 the caravan bearing Jacob and his family
to Joseph departed from al-'Arish, a village between Egypt and Canaan.-*

The technique was early established and must have had its origins in
haggadic exegesis. 5 Kalbi was moderate, if not quite restrained; in Muqa-
til's commentary to Svrat al Kahf no one and nothing remained anonymous.
I have adduced one example of the latter's employment of the device
(ta'yinltasmiya): to link two otherwise unrelated narratives by naming the
protagonist of each Yamlikha. 6 It would appear from the primitive treat-
ment of vague and anonymous references in scripture that these were

1 Speyer, Erzahlungen, 217; Kalbi, Tafsir, MS Ayasofya 118, i34 v ; cf. Zamakhshari,
Kashshdf ii, 494 ad loc, who also mentioned Reuben and Shimeon as possibilities; see
Schapiro, Elemente, 64-7.

2 Tafsir, i28 v , i35 r ; the Quranic conflation of the roles of Reuben and Judah is reflected
in Kalbi's exegesis, cf. Genesis 37: 21-2, 26-7.

3 Tafsir, iz8 v , I33 r_v , 128", i2Q r , I2Q V , 130", 135', respectively.
+ Tafsir, i2Q r , I30 r , 133*, I3Z V , I35 r , respectively.

5 See Goldziher, Richttmgen, 289-98.

6 See above, p. 128; that name achieved a degree of general usefulness in the interpre-
tation of Quranic narrative, cf. SuyutI, Itqdn iv, 86-7: variant Tamlikha; and Zamakh-
shari, Kashshdf ii, 720: the reference is Q. 37: 51.

13°

QURANIC STUDIES

regarded initially as lapses, or at least as inadequacies requiring the simple
corrective measure of amplification. With the elaboration of exegetical
method such vagueness of reference (mubham) was seen not only to be
intentional but also evidence of rhetorically sophisticated prose. That much
is clear from Suyuti's synoptic description of the subject, in which he set
out seven reasons for that phenomenon jn scripture: identification (ta' yin)
was unnecessary if the matter in question had been elucidated elsewhere in
the Qur'an, if it was too generally known to require such, or if there was
no particular value in closer specification; identification was undesirable
if the purpose of the mubham had been to attract attention by (partial)
concealment, or to emphasize general by excluding specific application,
or to achieve the effect of praise or of contempt by allusion rather than
direct designation. 1 Now, Muqatil and Kalbi were hardly concerned with
such nice distinctions, nor could it be said that their respective applica-
tions of ta'yin clarify in any way the scriptural passages so treated. But the
quality of the narrative was enhanced thereby and particularly, I suspect,
for the purpose of oral delivery.

Several elements in Kalbi's supplementary material to the Joseph story
derive not from the text of Genesis but from the Biblical tradition in a
wider sense. For example, ad Q. 12: 24 three interpretations were offered: 2

JISjj Ajjf ajj*> isij JISjj JUju jJ* Ajj (wi|o*/ajj jjU^J ^Ij o\ ^ ji

l$J ^ Asj <jU^ ^Tj o\ The first and third of these, which merely

paraphrase the Quranic locution, may be regarded as symbolic of the
author's concern to offer, wherever possible, more than one explanation.
The second: 'he saw the image of his father', draws upon an older and
well-attested tradition (cf . Genesis Rabba 87, 9 HXT TON pip'tt) and is
the only one that actually interprets the scriptural term burhan by ad-
ducing the gloss sura, and thus isolating the notion of 'manifestation'. 3
Ad 12: 67 Kalbi explained Jacob's advice to his sons to enter Egypt not
by one but by several gates as the father's fear that their striking beauty
could attract the evil eye: ^1 ^ l-jjaxi ( »$.U JU. <j£j, with which
may be compared Genesis Rabba 91,2 DdVd na»n Vxi im nnD3 lD»n
fSH , 1DS *rnn That kind of haggadic accretion, which must be

distinguished from material of strictly Biblical origin, was not limited to
the writings of exegetes like Kalbi and Muqatil, suspect in the judgement
of later generations for their undisciplined employment of Jewish material.*

1 Itqdn, naw' 70: mubhamdt al-Qur'dn iv, 79-100.
1 Tafsir, MS Ayasofya 118, 130'.

3 See Speyer, Erzahlungen, 201-3; Geiger, Was hat Mohammed, 139-40; Schapiro,
Elemente, 40-1; Rabin, Qumran, 113 n. 5; the standard explanation was burhdn:dydt,
see Suyuti, Itqdn i, 115.

4 See Speyer, op. cit. 214; Geiger, op. cit. 144-5.

s See Suyuti, Itqdn iv, 207-9; Goldziher, Richtungen, 58-60, 87, 112.

PRINCIPLES OF EXEGESIS

137

• t It is also found in the work of their contemporary, Sufyan Thawri, to
whom that stigma did not attach.

The Quranic exegesis of Sufyan (d. 161/778), parts of which have for
many years been known from citations in later writers, is extant in a
unique manuscript at Rampur. 1 The work, contained in eighteen folios,
consists of somewhat disjointed observations on forty-nine suras (from
Baqara to Tur, missing out Dukhdn and Muhammad) in the order of the
canonical text, though the internal sequence of verses is not that of the
canon. The fragmentary and uneven character of the work may be no more
than an accident and the compilation merely an aggregate of Sufyan's
opinions extracted from later works. That assessment will not, however,
explain the internal order of comment nor the quality of the explicative
elements themselves. It is those which, in the absence of a narrative
framework of the sort encountered in Muqatil, Ibn Ishaq, and Kalbi,
require particular scrutiny. In Surat Yusuf, for example, the symbolism
of Joseph's dream (Q. 12: 4) was interpreted 'his parents and his brothers'
and, alternatively, 'his father, his brothers, and his aunt', taking into
account Rachel's death before that event. 2 Although Kalbi had not made
explicit his knowledge of this fact until he reached verse 99, it would, I
think, be an error to assume that Rachel's earlier death was not generally
known to the exegetes as well as to the narrators of both the Quranic and
Biblical versions of the story. 3 At Q. 12: 24 burhdn was interpreted by
Sufyan as the figure of Jacob; at 12: 67 it was Jacob's fear of the evil eye
which prompted the warning to his sons; and at 12: 88 the locution bidaa
muzjat was glossed both 'little money' and 'butter, wool' (sic), reflecting
the much more detailed inventory of commodities adduced by Kalbi
to make up the gifts brought to Egypt by Joseph's brothers on their third
(sic) visit, a description very likely inspired by that of Genesis 43: 11-12
and one that became a stock item of the exegetical tradition. 4 Very oc-
casionally Sufyan is more informative than Kalbi, as at Q. 12 : 77, where the
cryptic 'if he has stolen then a brother of his stole before him' was inter-
preted 'Joseph had stolen their gods', exhibiting a confusion between
Joseph and Rachel which with very few exceptions persisted in Muslim
exegesis to this verse. Kalbi has merely an inconclusive reference to
Joseph, from which it is impossible to say whether the coat of many colours
or Laban's idols were intended. 5

1 GAS i, 518-19; ed. Imtiyaz 'AH 'Arshi.

2 Tafsir, 95- 1 07 ; the editor has rearranged the material in canonical order and indicated
the manuscript sequence in his numbering of the separate entries.

3 Sufyan, Tafsir, 95; Kalbi, Tafsir, 135"; pace Speyer, op. cit. 194 with reference to
Genesis 37: 10 and 44: 20; cf. Geiger, op. cit. 147—8.

4 Sufyan, Tafsir, 98 (adducing two traditions), 102, 104, respectively; Kalbi, Tafsir,
i3S r ; Zamakhshari, Kashshdf ii, 500 ad Q. 12: 88.

5 Tafsir, 103; Kalbi, Tafsir, 134'; cf. Speyer, Erzahlungen, 215-16; Geiger, Was hat
Mohammed, 145.

138

QURANIC STUDIES

It will be clear from these few examples that both the range and the
quality of Sufyan's glosses may justifiably be compared with those
of Kalbi. There is a shared tendency to transmit more than one inter-
pretation of a Quranic locution and, similarly, a concern with variae
lectiones. For example, ad Q. 12: 31 Kalbi commented oxxlj
gyi'^l (jjh Im*^ o! J of j o^Ji^-. ol J 6J W 1 * J ^*'J ^ eavm g °P en
the option between cushion and citrus (etrog: some manuscripts read
utrunj), whereas Sufyan, also adducing a variant, restricted the choice to

one between foodstuffs \&] J ^ J 15 Jjt, U« ^t- ^ Uu*.

gjyifl Jli ^ j-UkSI Jl» l^j- 1 Since the entertainment

provided by Potiphar's wife clearly drew upon Rabbinic tradition, it
may reasonably be suggested that Sufyan's represents the earlier choice
of interpretations, and that the proposal to read 'cushions' (muttaka'an:
wasayid) exhibits yet another instance- of redactional intervention in
the transmission of Kalbi's commentary. 2 To Muqatil's glosses, too, those
of Sufyan may be compared, as for instance ad Q. 18: 46 and 19: 76
al-bdqiydt al-salihdt were interpreted as the five (ritual) prayers, that is,
as salat rather than as dud, but the equivalence works: prayers is a
common ground. 3 Thus, from this sampling of its ingredients the ex-
egesis of Sufyan can be described as belonging to the haggadic type.
There are none the less some conspicuous lacunae: not only is Surat
al-Kahf shorn of its traditional narrative framework, passages normally
pegs for extensive and varied anecdote, like the opening verses of suras
17 and 30, are here given no attention whatever. Omissions such as
these, like absence of comment for Dukhdn and Muhammad in the
Rampur manuscript, are difficult to explain, even if that document were
to be no more than an extrapolation of Sufyan's utterances from later
writers (e.g. 'Abd al-Razzaq, Tabari, RazI), rather than the fragment
of an independent work. External evidence, such as it is, appears to lend
support to the latter alternative, though I am unable to accept without
reservation the remark of Ibn Abl Hatim that Sufyan disapproved of those
who, like Kalbi, commented on the entire text of a sura even where there
were no problems to be solved. 4 Sufyan's glosses are of the quality charac-
teristic of an original narrative framework and are virtually interchangeable
with those of his contemporaries, here designated haggadic. But it may be
recalled that for posterity the reputation of Sufyan, unlike those of Muqa-
til, Kalbi, and Ibn Ishaq, remained unblemished.
Problems of transmission and redaction history are notoriously complex.

1 Kalbi, Tafsir, 130*; Sufyan, Tafsir, 100.

2 Cf. Speyer, op. cit. 205-6; Geiger, op. cit. 140-2; and see above, pp. 132-3.

3 Tafsir, 136, 147; cf. above, p. 133, for Muqatil, Tafsir, ijo T ad Q. 18: 46.

4 Editor's introduction, 33-8; Taqdima, 79, cited in introduction, 16.

PRINCIPLES OF EXEGESIS

139

v A parallel to the relationship between the Rampur manuscript and those
dicta ascribed in later works to Sufyan (as set out in the editor's detailed
apparatus) could probably be found in a comparison of opinions attributed
to Mujahid b. Jabr (d. 104/722) and adduced by Tabari (d. 311/923) with
the Cairo manuscript of Mujahid's Tafsir. 1 Remarkable indeed is the use by
Tabari of Mujahid to support what Goldziher described as 'rationali-
stische Koranauslegung'. 2 In the light of that argument it would be of
considerable value to examine Mujahid's methods in the context of his own
work. An obstacle to the kind of comparison suggested is posed by the
practice, widespread in later exegetical writings, of introducing minority,
dissenting, and unpopular interpretations anonymously. 3 The technique
may be illustrated with reference to Q. 12: 31, noticed above in a compari-
son of Kalbi with Sufyan: now, Zamakhsharl (d. 538/1143) offered for the
enigmatic muttakan\muttakcL an the following possibilities: place in
which to recline (literal), place in which to eat (metaphorical), food
whole or sliced (metonymical), and citrus (caique of Hebrew etrog) in that
sequence and with several orthographical variants, some attributed
others introduced wa-qtla/zoa-quri'a. Out of context as it were, and anony-
mous, the etrog etymology neither served the same purpose nor produced
the same effect as when situated within the haggadic framework. 4 Zamakh-
shari's work presupposed both methods and results of the haggadic,
halakhic, and masoretic types. Quranic interpretation had long since
achieved the status of normative discipline and the exegete was free to
select from the tradition those elements best suited to his purpose and,
moreover, to arrange them according to one of a large number of priorities.
The original aim and/or significance of a gloss might be accidentally
overlooked or intentionally discarded, its typical context thus ultimately
forgotten.

Drawing upon these elementary observations I would submit that the
attempt to extrapolate from later works those of earlier authorities is bound
to produce both incomplete and inaccurate results. It was, not surprisingly,
Wellhausen who first applied the (Pentateuchal) Urkundenhypothese to
Arabic literature. The end of that exercise, available in both his Skizzen
und Vorarbeiten VI (1899) and Das arabische Reich und sein Sturz (1902),
was to isolate regional and partisan tendencies in Tabari's monumental
history of the Islamic world. While it would be ungracious not to acknow-
ledge that this was an interesting and valuable experiment, one will be chary
of concluding from it that Tabari's primary sources have been or can be
recovered in a form at all close to their original state. To assert the contrary
would imply that Tabari's work is merely a compilation, exhibiting little

1 GAS i, 29 ; see Horst, 'Zur Uberlieferung', 295-8, 307.

2 Richtungen, 88, 107-10. 3 See above, pp. 120-1 for an example ad Q. 7: 158.
4 Kashshdf ii, 462-4 ad loc.

140

QURANIC STUDIES

or no trace of the writer's craft. Such I find impossible to accept and in
illustration mention the artfully composed account of Ibn Ash'ath's
revolt, allegedly transmitted from Abu Mikhnaf (d. 157/774) and consisting
almost entirely of ayyam motifs constructed round a fluctuating employ-
ment of first-person narrative. 1 For application of the Urkundenhypothese,
and of other principles of Biblical literary (documentary) criticism, it is
well to distinguish between questions about origins (chronologische Anset-
zung) and those designed to isolate parallel, divergent, and conflicting
strands within the literary tradition (Herausschdlung der Fdden). 2

From the point of view of chronology, the development of Muslim
exegetical literature envisaged here required a span of approximately a
century and a half, from Muqatil (d. 150/767) to Ibn Qutayba (d. 276/889).
Within that period the principles of exegesis were evolved and perfected,
and it would not be too much to say that thereafter few, if any, methodo-
logical innovations were introduced. For isolation and description of its
components the selection of criteria is a matter requiring the greatest care.
In his analysis of Tabari's history Wellhausen employed exclusively the
factor of ascription, a choice rendered deceptively attractive by Tabari's
fairly consistent use of chains of transmission. But even when supported by a
highly differentiated nomenclature for the modalities of transmission, ascrip-
tion can be remarkably unstable. 3 It is on the one hand vitiated by internal
contradiction (as in dicta attributed to authorities like Ibn 'Abbas and the
prophet), and on the other attenuated by anonymity {wa-qilajwa-quri'a).
Ascription is also arbitrary: biographical information on the exegetes is
found exclusively in literature composed to impugn or to vindicate
(jarh wa-ta'dit) or to assess relative merit (tabaqat), and as such constitutes
merely a pseudo-historical projection of the acceptance or dismissal of
their views. For these reasons I have thought it best to ignore, or at least
to discount, ascription and, by concentrating on the elements of explica-
tion both in and out of context, to isolate and identify methodological
devices which can be recognized without resort to biographical data.

In the four examples of exegetical writing so far considered I have
underlined the centrality of the narratio, which is normally accompanied
by or appears itself to generate a number of typical devices. Many of

1 Annates ii, 1064-77; cf- BSOAS xxxiii (1970) 616; on Tabari's manipulation of
tradition cf. Birkeland, The Lord Guideth, 9, 10, 16, 22, 29, 40-1; and Widengren, 'Oral
tradition', 244-58.

2 See Eissfeldt, Einleitung, 185-216; and above, I pp. 16-17 on Wellhausen's Reste;
for the tyranny of the 'literary critical' method see also the observations of Mowinckel,
'Psalm criticism', 13-33, and Richter, Exegese, 66-7, 120-2, 145-52.

3 e.g. Abbott, SALP i, 5-31, ii, 5-83, 106-13; Sezgin, GAS i, 53-84, 237-56; despite
careful and often illuminating analysis of technical terminology, the studies of both
authors suffer, in my opinion, from an ingenuous acceptance of the isndd apparatus, but
represent at the same time a not altogether unexpected reaction to the work of Goldziher
and Schacht.

PRINCIPLES OF EXEGESIS

141

», those are means of sustaining continuity: prolepsis and cross-reference,
repetition/subsumption, variable connectives, interpolation and para-
phrase. Others, like segmentation, supercommentary, apostrophe/paren-
thesis, and recurrence of the minimal units, would seem to, but in fact
do not, interrupt the narrative flow. In addition to those stylistic devices,
narrative elements such as anecdote, prophetical tradition, identification
of the vague and ambiguous [ta'yin al-mubham), and description of the
occasions of revelation (asbab al-nuzul) are present in varying quantity,
always sufficient to identify the haggadic type. Even where the narratio
itself is absent, as in the Rampur manuscript of Sufyan's exegesis, the
presence of a number of these elements makes possible identification of the
type. Of the twelve explicative elements proposed as criteria for a descrip-
tive analysis of exegetical literature, three at least may be regarded as
typically haggadic: anecdote, prophetical tradition, and identification.
Others of them are also found there, but in a relation to the type which
I should call accidental rather than essential. Some, like^ poetic loci pro-
bantes and variae lectiones (and the related alternative explanation), are
clearly intrusive, not least owing to their disruptive effect upon the
narratio.

A special case is description of the occasion of revelation, characteristic
of halakhic exegesis but present in underdeveloped form in the haggadic
type. 1 By that I mean merely that the essential function of the sabab
al-nuzul (or tanzil), which was to establish a chronology of revelation, is
not evident in haggadic exegesis. There the value of the device is ex-
clusively anecdotal, and may provide the narrative framework for an
extended interpretation, either of a whole sura as in Muqatil or of frag-
ments of many suras as in Ibn Ishaq. In that sense of course it could be
argued that the entire narratio functions as tanzil. The formulae usually
employed to introduce the device are nazalat (hddhihi 'l-aya) fifuldn and
wa-dhalika hinajamtahu qdla fulan, often accompanied by an anecdote to
provide background or local colour. But almost never in haggadic exegesis
is the tanzil qualified by alternative explanations of the circumstances of
its revelation or followed by a discussion of its juridical significance. Even
where interpretation was adapted to an over-all narrative structure, as in
MuqatiPs treatment of Surat al-Kahf, separate occasions of revelation may
be adduced. An example may be seen ad Q. 18: 28, where 'you desire the
vanity of this world' was said to have been revealed in reproach of the
extreme vanity of TJyayna b. Hisn during an interview with Muhammad,
into the account of which was inserted an allusion to the social inequality
obtaining between Arab and mawld. Since 'Uyayna's vanity was pro-
verbial, it may well have been not that but the social motif which prompted
inclusion of this particular tanzil. 2 An occasion of revelation may be
1 See below, pp. 177-9. 2 Tafsir, i6g r .

142

QURANIC STUDIES

gratuitously specified, as in Kalbi ad Q. 12: 7, where the passage 'there
are lessons for those who ask' is followed by 'this verse was revealed with
reference to a story of the Jews'. The purpose of that disclosure in a
context exclusively Israelite is not immediately clear, but since it is the
only example of tanzil in Kalbi's treatment of Sural Yusuf, I am tempted
to regard it as merely formal acknowledgement of a methodological prin-
ciple incorporated into a later redaction and foreign to the original version
of Kalbi's exegesis. 1 The frequency of tanzil in haggadic exegesis varies,
and in Sufyan is concentrated in suras containing in fact a high degree of
halakhic content. But even there it is sporadic and unpredictable, and its
value merely anecdotal, e.g. ad Q. 2: 125, 2: 164, and 2: 186. Exceptions
are seldom, as at Q. 2 : 143-4, where the change of qibla from Jerusalem to
Mecca was dated. 2 For later theorists the expression 'this verse was revealed
about . . .' contained a significant ambiguity, by means of which it became
possible to distinguish between a cause (sabab) of revelation and a report
(khabar) about it. According to that distinction tanzil of the sort found in
haggadic exegesis could be relegated to the status of khabar*

The most flagrantly intrusive of the explicative elements found in
haggadic exegesis is poetry adduced to explain Quranic lexica. It is not
present in Muqatil, Kalbi, or Sufyan, and instances in the Sira may be
attributed to the editorial intervention of Ibn Hisham. That method of
interpretation belonged to the masorah and was intimately related to the
contemporary development of techniques for the transmission of poetic
texts. 4 Solutions to lexical problems were sought by the haggadists
within the vocabulary of scripture itself, by recourse to a crude but
apparently effective kind of textual analogy. For that device Muqatil
employed the term nazir, but also mithl and shabah, and occasionally the
particle ka. With the exception of nazir, these terms designating analogue
appear also, though less frequentiy, in the work of Kalbi and Sufyan.
Together with the term toajh ('reference' as contrasted with 'information'),
these formed a technical vocabulary for the distributional analysis of
meaning in scripture. 5 Closely related thereto is a series of concepts attach-
ing to the terms mushtabih and mutashabih which, with wujuh and naza'ir,
were not fully developed before elaboration of the masorah. 6 I have there-
fore deferred discussion of them, though it may be observed that as with
tanzil, concern with recurrent and even crucial terms in scripture is found
in its earliest stage of development in the haggadic literature.

Another kind of lexical treatment was that accorded hap ax legomena

and words conceded to be of foreign origin, as when Muqatil ad Q. 18: 31

explained istabraq as a Persian word {lughat farts) for brocade, or ad 18:

1 Tafsir, 128". 2 Tafsir, 9, 14, 17, 11, respectively.

3 Cf. SuyutI, Itqdn i, 90; see above, I pp. 41—3, and below, pp. 177-8.
+ See above, III pp. 94-8; cf. BSOAS xxxiii (1970) 390, 616.

5 See above, III pp. 99-100. 6 See below, pp. 208-16.

PRINCIPLES OF EXEGESIS

143

s r 107 fir daws as Latin/Greek (lughat al-rum) for walled garden. 1 That
procedure, found also in Sufyan, e.g. for qistas in Q. 17: 35,2 but not in
Kattri, may be of some value in dating efforts to derive the scriptural lexi-
con from exclusively Arabic origins, a process which, though associated
with the name of Ibn 'Abbas, was almost certainly the product of maso-
retic exegesis. 3 Like the variants (both textual and explicative) found in
the work of Kalbi and Sufyan (but not in that of Muqatil and Ibn Ishaq),
concern with the lexicon might be thought to presuppose a standard if not
quite ne varietur text. The earliest method for dealing with basic (crucial)
words/notions appears to have been simple paraphrase, as in Kalbi ad

Q. 12: 87 -mi ju*j ^/.uii £jj ij~Lu

The elements here briefly described and characterized as intrusive
within the haggadic type belong, with one exception, to the interpretative
paraphernalia of the masoretes. The exception, mention of the occasion
of revelation, is essentially halakhic. Now, isolation of these devices at the
haggadic level is difficult, owing to the virtual absence of fixed technical
terminology. Save for nazir (analogue), Muqatil employed in his Tafsir
only two other terms which could be described as of more or less rigorous
technical application: isti'naf (juncture) and taqdim (hyperbaton). Even that
minimal evidence of technical vocabulary cannot be found in the works
of his contemporaries, and it is not quite impossible that these terms, too,
could be regarded as intrusive. While there is not for Muqatil evidence
of redactional activity of the kind available for Ibn Ishaq, the following

entry a d Q. 18: 22 deserves notice: 5 j»^>SGI _jl L Ujj

j^JU oJUfej l J^i\ d& JWI jlj jljJI ajj& e-l Jli i_Jliu ^LjJI y\ JlS

i_J50I jlxc. It may first of all be remarked that this kind of close
philological treatment is not at all typical of Muqatil's exegesis. Insertion
of a conjunction into the last only of a series of distributive enumerations
was not the sort of problem which interested the author of that work, and
one is thus not surprised, apart from the anachronism, to find cited
the grammarians Farra' (d. 207/822) and Tha'lab (d. 291/904). Curiously,
the explanation is not found in the commentary of Farra', for which Tha'lab
was principal ream, despite concern there with grammatical niceties. 6 Nor
do the several examples of hyperbaton adduced by Muqatil for Surat al-
Kahf(i.e. ad verses 6, 10,21, 25, 31 : all instances of an indefinite accusative/
adverbial, shifted from a logical to apausal position) figure in Farra 's
treatment of those verses, though these were admittedly not referred to the
later authority. Now, the recension in which Muqatil's exegesis was

' Tafsir, 169*, 173'. 3 Tafsir, 131. 3 See below, pp. 218-19.

♦ Tafsir, 135 1 . 5 Tafsir, i68 v .

6 Ma'am 'l-Qur'an ii, 138 top, ad loc.; GAL i, 116, Suppl. i, 178 (Farra'); GAL i,
118, Suppl. i, 181 (Tha'lab).

144

QURANIC STUDIES

transmitted is that of Hudhayl b. Habib, who can hardly have been respon-
sible for mention of Tha'lab and probably not even of Farra'. 1 But external
evidence of that kind is after all secondary, since the intrusive character
of the commentary to Q. 18: 22 can be established on structural grounds.

Similarly, structural analysis of Kalbl's exegesis provokes some doubt
about the authenticity of variae lectiones attributed to the author. 2 There
the elliptic phraseology might seem to presuppose the detailed discussions
of textual variants found in masoretic authorities, e.g. Farra'. Because of
its fragmentary state Sufyan's exegesis is more difficult to assess. As in
other examples of haggadic commentary the explicative elements there
consist mostly of paraphrastic equivalents, but shorn of an over-all
narrative structure which could have provided stylistic uniformity. Refer-
ences in the body of the Tafsir to Sufyan's rawi, Abu Hudhayfa (d. 240/
854), e.g. at Q. 2: 297 and 36: 12, supply a date which might explain the
occasional appearance there of masoretic (varia lectio) and halakhic
(tanzil) elements. 3 It must, I think, be recognized that extant recensions of
exegetical writing here designated haggadic, despite biographical informa-
tion on its putative authors, are not earlier than the date proposed to mark
the beginnings of Arabic literature, namely 200/815. For the relationship
between (canonical) text and commentary the implications of that acknow-
ledgement will be obvious: original distinctions have been blurred by
redactional activity. The fact itself of literary transmission, moreover, will
have contributed to a degree of stylistic and methodological uniformity
throughout the range of exegetical literature that makes difficult, if not
quite impossible, description of the Sitz im Leben of any of its types.

At the beginning of Sitrat al-Riim, Muqatil told the story of a wager
between Abu Bakr and Quraysh on the number of years to pass before a
Byzantine victory over the Persians would wipe out the humiliation of their
defeat by the latter mentioned in scripture. Besides emphasizing the
reading ghulibat and glossing consistently the scriptural muminun as
musUmun, the story combines an example of Quranic prognostication (akh-
bar al-ghayb) with a neat connection between events in the Hijaz and
the wider world. 4 The primary motif, a natural alliance between Muham-
mad's followers and the Byzantines (both being 'people of the book')
against his opponents and the Persians (both being idolaters), became
a constant in Quranic exegesis and a 'fact' of oriental history. 3 The circular
argumentation underlying that process is graphically illustrated by the
manner in which Ahrens drew upon Wellhausen's assertion (itself ap-
parently an inference from the haggadic interpretation of Q. 30: 1—4)

• GAS i, 37: 'dernoch 190/805 lebte'. 2 See above, pp. 132-3.
3 Tafsir, zz, 208; GAS i, 41, but cf. the editor's introduction, Tafsir, 38, adducing the

dates 220/835 and 226/841.

* Tafsir, 230 v ad Q. 30: 1-4; see above, II pp. 69-70.

5 e.g. Ibn Sa'd, Tabaqat ii, 24; and references in Kister, 'Al-rJIra', 144 nn. 2-3.

PRINCIPLES OF EXEGESIS

145

^ that the Jews in Arabia (hence opponents of Muhammad) had tradi-
tionally ( !) sided with Persia against Byzantium, to prove, conversely,
that Islam was influenced in its development by the prophet's sympathetic
attitude to Christianity. 1 I have mentioned the absence of comment to this
passage in the Rampur manuscript of Sufyan; Kalbi alluded to the wager
and, like Muqatil, linked the eventual Byzantine victory with the Muslim
one at Badr. Thus, what for the exegetes could only be regarded as vati-
cinatio ex eventu furnished anecdotal material both entertaining and edi-
fying. 2

Narrative ingredients such as that may be described as pseudo-historical
digressions. 3 Other kinds are also found, e.g. the prophetical hadiih adduced
by Muqatil in which was described the barrier erected against the depre-
dations of God and Magog, or by Sufyan, where three gentlemen, physi-
cally ample but intellectually feeble, discussed in an ingenuous way the
kind of conversation God might be expected to overhear. 4 Each story
provides a very literal, almost tactile, realization of the verse in question.
Related to these features is the apostrophe (parenthesis), in which the
exegete addressed his audience by paraphrasing and amplifying a scriptural

locution, as in Muqatil ad Q. 18: 69 Jlii/ljjU* <U)1 cLi <jl JJbiZ-, J IS
Ajji jh, pjj jy^j. As JJUu,* or Kalbi, ad Q. 12: 99 I^Lol JlSj
tj*Jlj jJbJI aM\ tli Jij/^jjuT «UII &L2 oj Like most of

the haggadic techniques examined here, of which the primary example
is the narratio itself, digressions such as those do not so much explain
scripture as render it familiar. No device could, after all, have been
more appropriate to the task of making familiar than provision of names
(ta'yin) for the anonymous. 7 The situation into which such procedures
could be insinuated might be thought a very informal one indeed.

I have suggested that regular expression of connectives and employment
of supercommentary could indicate oral delivery; similarly, 'stage direc-
tions' following qala, as well as serial repetition and circular explication,
would seem unnecessary in a text designed to be read rather than heard. 8
That the evidence both of style and of content should point to the popular
sermon as Sitz im Leben of haggadic exegesis is hardly surprising. 9 For the

1 'Christliches im Qoran', 148 ; Reste, 236; what may have been the source of this align-
ment is attested in the history of Palestinian Jewry, see Mann, Fatimid Caliphs i, 42 citing
Graetz.

2 See above, p. 138; Tafsir, 228 r ~ v ; worthy of remark is the unquestioning acceptance
by the haggadists of the reading gkulibat, cf. GdQ i, 149-50; Goldziher, Richtungen, 18-19.

3 The example is not isolated, cf. Goldziher, Richtungen, 58-61, especially on Muqatil
ad Q. 17: 60.

4 See above, p. 133; Tafsir, 226-7 adQ.41: 22.

5 Tafsir, MS H. Husnu 17, 171*.

6 Tafsir, MS Ayasofya 118, 135*. 7 See above, pp. 135-6.
8 See above, pp. 128-31. 9 See above, I pp. 46-52, esp. pp. 48-9.

146

QURANIC STUDIES

history of Arabic literature a long period of oral composition and trans-
mission, or possibly of oral delivery from notes, is commonly supposed to
have preceded the redaction of more or less fixed texts. It is the chronology
of that process which eludes satisfactory description. Two points in this
respect deserve some attention. First, literature exhibiting the haggadic
type is preserved only in recensions dating from the third/ninth century.
The presence there of what I have described as intrusive elements, though
indicative of an incipient sophistication of exegetical method, is not of such
dimension as to distort the basic character of the type, which remained
recognizable and continued presumably to perform some useful function
in the community. Second, it seems clear from the manner in which
textual problems were treated, or ignored, that development of the haggadic
type preceded in time the refinement of method characteristic of masoretic
exegesis. An alternative would be to suppose that the two types developed
in mutual isolation and perhaps simultaneously. That view could derive
some support from scattered attempts to perpetuate the haggadic type,
for example, by Dlnawari (d. 308/920) or Qummi (d. 309/921). 1 The former
work, a nearly verbatim reproduction of Kalbi's commentary and like the
latter transmitted on the authority of Ibn 'Abbas, is hardly conceivable at
a time when the writings certainly of Ibn Qutayba (d. 276/889) and prob-
ably of Tabarl (d. 311/923) were available. It may be that Dlnawari
considered the work of his contemporaries (quite correctly) inappropriate
to the pulpit, but if that were so the work of Kalbl himself or of Muqatil
could have been used. Qummi's Tafsir (here the title could be authentic),
on the other hand, may have been composed to meet a different need.
Purportedly derived from the authority of Ja'far al-Sadiq and of his father,
the commentary consists entirely of haggadic elements applied to sectarian
theology and displays, curiously, very little in common with the allegorical
exegesis contained in writing attributed to Ja'far. 2 Lexical explanation is
based on paraphrastic equivalence, textual emendation on 'Alid symbolism,
poetic shawdhtd are minimal, and the use of narratio abundant and in con-
formity with haggadic practice. An interesting variation, in his account of
the 'rabbinical' test of prophethood, was location of the rabbis at Najran. 3
Perpetuation of the haggadic type might thus be ascribed to the survival
of its traditional function within the Muslim community; it can hardly be
explained within the framework of literary history. Formally haggadic
elements in the exegesis of Tabari and his successors were functionally of
another order, and had been adapted to a different set of priorities.

Source materials for the popular sermon and other forms of public

1 GAS i, 42 (Dlnawari), MS Ayasofya 221-2; GAS i, 45-6 (Qummi).

2 Cf. GdQ ii, 180; Goldziher, Richtungen, 379 ff; GAS i, 528-31 (Ja'far), and see
below, pp. 245-6.

3 Qummi, Tafsir ii, 31-2 ad Q. 18: 9.

PRINCIPLES OF EXEGESIS

147

\ oratory are remarkably unstable. An apposite illustration is the khutba
ascribed by Jahiz to 'Abdallah b. Mas'ud, adduced verbatim by WaqidI,
who attributed it to the prophet during his expedition to Tabuk. 1 Literary
form can be of assistance in determining date if not always authenticity.
It is perhaps not without interest to observe that the parallel employment
of elative constructions in that khutba is found also in the laudatio of Umm
Ma'bad, allegedly composed during the prophet's hijra, and in the address
to Quraysh by Nadr. b. Harith on the appearance in their midst of Muham-
mad. 2 The earliest examples of oratory seem to have been characterized by
synonymous or synthetic parallelism (mutabaqa) rather than by rhymed
prose (saj') f a conclusion supported by the form of speeches in the ayyam
literature. 3 It is, on the other hand, quite impossible on the basis of such
material to date the appearance of rhymed prose or, more important, to
infer reasons for abstention from that particular form. 4 Now, a historical
development described in terms of evolution from pre-Islamic khatib to
Islamic qdss might be thought to reflect the argument in favour of fasahat
al-jahiliyya, were it not invariably accompanied by a portrait of the
popular preacher as degenerate and irresponsible purveyor of fable. 5 That
the designation qdss became an epithet of abuse may have been a con-
sequence in part of the fact that he remained a 'popular preacher' on the
periphery of the religious establishment. Opprobrium might thus reflect as
much functional eccentricity as doctrinal irregularity. 6

Much if not all of his material, however, is found in the writings of the
haggadic exegetes. It may not be irrelevant to note that the Quranic scrolls
from Damascus, composed certainly for the purpose of private and possibly
communal devotion, contain almost exclusively scriptural passages which
could, and did, generate haggadic material, e.g. prophetology, eschatology,
and paraenesis. 7 To perceive in the origins of Arabic literary prose a com-
bination of public oratory and elaboration of 'predication coranique' is
undoubtedly sound. 8 Elaboration must of course be understood as inter-
pretation, of which in this context the typical variety was that represented
by the aetiological legend, like those related of Abraha's elephant and
Muhammad's nocturnal journey. 9 The narratio was both didactic and

1 Jahi?, Al-Baydn wal-tabyin ii, 52; Waqidi, Kitab al-Maghdzi, 1016.

2 See Fischer, 'Umm Ma'bad-Legende', 318-27; Ibn Hisham, Sira I, 299-300; cf.
also Stetter, Topoi und Schemata, 42, 45-6.

3 See Caskel, 'Aijam al-'Arab', 45-6.

* Pace Goldhizer, Abhandlungen i, 57-76, esp. 67-8; cf. Fischer, 'Umm Ma'bad-
Legende', 318 n. 1; BSOAS xxxiii (1970) 390; and see above, III pp. 1 16-17.

5 See above, III pp. 93-9; cf. Goldziher, Studien ii, 161-70; id. Richtungen, 58-61;
id. 'Chatib', 97-102; id. 'Neue Materialien', esp. 478-9.

6 See Pedersen, 'Islamic Preacher', 226-51; id. 'Criticism', 215-31.
' See Ory, 'Un nouveau type', 87-149, esp. 144-9.

8 See Blachere, Histoire jii, 717-36, 737-803, though I am unable to accept the author's
proposed chronology.

9 See above, I pp. 42-3, II pp. 67-70.

148

QURANIC STUDIES

entertaining, and anecdotal accreta appended to scriptural texts conformed
admirably to the pre-halakhic concept of pious and edifying tradition,
symbolized in the formula <u ^'ti^ jjSOj iju** l>jJU (poorly accredited
but of therapeutic value). 1 To the long and many-faceted process of Ge-
meindebildung which culminated in the canonical text of Muslim scripture,
the sermon (khutba) must have been central, as the instrument of both
transmission and explication of the prophetical logia. The role of Haggadah
was described by Zunz as that which most easily and naturally met similar
needs in the post-Exilic Jewish community. 2 The manner in which the
popular sermon and the popular preacher were eventually incorporated
into, or eliminated from, the orthodox establishment belongs to the internal
history of the religious community. The strictures of halakhic and maso-
retic exegesis did not of course preclude oral delivery, but probably limited
such to the lecture room. The requirements of a wider public were not for
that reason neglected.

2. DEUTUNGSBEDURFTIGKEIT

Concern with both hermeneutical value and grammatical form of revela-
tion could be justified by recourse to scripture itself, whose Deutungs-
bedurftigkeit was in more than one passage explicitly stated. 3 The related
but distinct processes of hermeneutical derivation and textual adjustment,
neither of which figured more than marginally (or intrusively) in the work
of the haggadists, were conveniently described by Vermes as 'applied' and
'pure' exegesis respectively. 4 Those labels are eminently practical, and
indicate functional value rather than methodological content of the
exegetical types. Related to the formation of the Islamic community, and
measured against the data of Arabic literary history, both kinds of exegetical
activity (or rather, hermeneutics and exegesis proper) represent phenomena
typologically distinct from the haggadic expression analysed in the pre-
ceding pages. These phenomena consist principally in the elaboration of
analogical method and in the concomitant acquisition of a technical
vocabulary. While elements of the latter can often be traced to scriptural
usage, the further semantic development of exegetical terminology usually
followed paths divergent from, and even contrary to, the rudimentary
associations of scriptural context. That such is so for the vocabulary of
Biblical exegesis was stated recently by Loewe and is amply demonstrated
in Bacher's lexicon. 5 If the elaboration of Rabbinic, and sectarian, exegesis

1 Goldziher, Studien ii, 154; cf.BSOAS xxxi (1968) 615.

2 Vortrage, 342-441, esp. 354-73; and Griinbaum, Sagenkunde, 1-54; cf. Elbogen,
Gottesdienst, 194-8; Seeligmann, 'Midraschexegese', esp. 176-81; Vermes, Scripture,
1-10, 67-126.

3 See above, III pp. 99-102. ♦ 'Bible and Midrash', in CHB i, 199-231.
s 'The "plain" meaning', 154; Bacher, Terminologie i-ii, passim.

PRINCIPLES OF EXEGESIS

149

can be shown to reflect the impingement of Hellenistic rhetorical tradition, 1
the corresponding evolution of Muslim exegetical terminology, closely
associated with the Jewish tradition, was even more complex. The emer-
gence, at the end of the third/ninth century, of rhetorical criticism appar-
ently derived from and certainly directed to works of profane literature
may be, and indeed has been, interpreted as evidence of Hellenistic in-
fluence upon the Arabic science of rhetoric. 2 That particular view is of
course only one of several possible. But whatever the ultimate source of
any given procedure or device, it is quite impossible to separate the de-
velopment of profane rhetoric from that of scriptural exegesis, at least in
any but the haggadic sense of that expression. However contrived and
exclusively theoretical the relation of Arabic eloquence to the word of God
might seem, the tendency to seek in scripture authority for the principles
of rhetoric was very real. 3

Symbolic of the alliance between the two disciplines was the use made of
Q. 3: 7 i_>h50l |»t oLSo*-, obT ax* <->bS3\ dLi& J^I ^jJI _yt>
&U^lj pUil <cl* 4jl£J U jjjtjzJ £ij (*4?jJ^ (j ^^Jl L»l» ol^jl.l." <
Xs. ^ JS" «u lliT djiyi j^LJI j jjpcJ J\j <UJ| "i\ aJujIj pl»u Lj -»Ljtj
vM^I IjJjJ L»J ^j- Commentary on this passage, unanimously

agreed to represent the point of departure for all scriptural exegesis,
itself exhibits a historical and typological spectrum of interpretative
method. The operative terms in the verse were seen to be three: muhkam,
mutashdbih, and umm al-kitab, each of which came to be assigned a dis-
tinguished if uneven semantic history. Concern at the haggadic level was
naturally with unitary definitions, and for Muqatil the ayat muhkamat
were those verses whose prescriptions were to be implemented, further
specified (or exemplified) as Q. 6: 151-3. Such were designated mother/
source of the book (umm al-kitdb:asl al-kitab) since they were not only
preserved with God(fi'l-lawhi 'l-mahfuz: sic, cf. Q. 85: 21-2) but also in
the scriptures of all peoples (sic).* Kalbl's view was not dissimilar, but
included the additional qualification that these verses set out permission
and prohibition and were ones which had not been abrogated (mubayyinat

bil-halal wal-haram lam tunsakh). 5 For both exegetes the ayat mutashabihat

/-> r> r->

were the four initial sigla J\j ^aJIj pJI found in thirteen suras,
and here of numerical and apocalyptic value, related to the taunts

1 Cf. Loewe, op. cit., 140-54; Daube, 'Rabbinic methods', 239-64; Gertner, 'Terms'

I- 37 ; id. 'Pharisaioi', 245-68.

2 See Taha rjusayn, 'La Rh&orique arabe', 3-24; Heinrichs, Arabische Dichtung,

II- 18, 105-70, with reference to von Grunebaum, 'Die aesthetischen Grundlagen der
arabischen Literatur', Kritik, 130-50.

3 See above, II pp. 79-80, and below, pp. 232-9.

♦ Tafsir, MS H. Hiisnvi 17, 35 T . 5 Tafsir, MS Ayasofya 118, 29'.

QURANIC STUDIES

of Muhammad's Jewish opposition. 1 That this kind of interpretation
served a useful, if limited, purpose may be clear from the preceding
observations on haggadic method. It could not, and did not, survive the
more exacting demands of halakhists and masoretes. Kalbi's reference for
muhkam to the principle of abrogation (naskh) necessitated a correlation
mutashabih : mansukk, implicitly ascribed to Ibn 'Abbas and extending con-
siderably application of the term mutashabih. In rather more detail, and
with explicit ascription to Ibn 'Abbas, was the introductory statement of
Abu 'Ubayd (d. 224/838) in his treatise on abrogation, for which Q. 3 : 7

must have seemed to the author an appropriate peg: i3*J& oUSs?tJI

Aj J^xj <u jyjj Uj -uUifj -UliJj 6j^j^,j. z By means of that aphor-
istic formulation, which did not add appreciatively to delimitation of a
technical vocabulary, scriptural material not of regulative content, in
the opinion at least of Abu 'Ubayd, was relegated to the status of muta-
shabih: 'the object of belief but not of conduct'. That functional cleavage
could hardly be of use for the masorah, and it is thus curious to
find repeated by Farra* the information in Muqatil and Kalbi, including
the allusion of the latter to abrogation. 3 The same material was also
adduced ad loc. by Zajjaj (d. 311/923) who, however, proposed a further
contrast: muhkam verses were immediately meaningful owing to their
straightforward/obvious expression (zahir bayyin), while mutashabih verses
in order to be understood required insight (nazr) and reflection (tadabburj
tadbir). Examples of the first category were the stories of the prophets and
of creation (sic), of the second the claims for the fact of resurrection. 4
Explicit reference to the 'plain meaning' of scripture might seem arbitrary,
if not quite insidious, in a work where the muhkamat were also subjected
to exegesis. Even a theoretical postulate that the muhkamat were imme-
diately clear (zcadik tnubin) was rejected by Maturidi (d. 333/944) in his
detailed survey of the several traditions relating to Q. 3 : 7.* A series of
contrasting pairs was set out: the meaning of muhkamat could be
rationally apprehended (fil-aql baydnuhu), that of ntutashdbihdt only by
recourse to authoritative tradition (bi-ma'rifat al-sam'); muhkamat were
verses of regulative content (ahkdm), while knowledge of the mutashabihdt
was not even necessary (laysa bil-nds hdja ila 'l-'ilm bihi); muhkamat
were abrogating, mutashabihdt abrogated ; muhkamat could be understood

1 Cf. GdQ ii, 68-78; and above, II p. 64; Muqatil and Kalbi are too allusive: the
entire anecdote on the expected duration of Muhammad's power is retailed in Suyuti,
Itqdn Hi, 25-6.

x GAS i, 48; MS Ahmet III 143, 3'.

3 Ma'ani 'l-Qur'an, MS Nurosmaniye 459, 29 v ad loc.

* GAS i, 49; MS Nurosmaniye 115, 67": entitled 'IrdbjMa'dnt 'l-Qur'dn.

5 GAS i, 49, 604-6; Taviildt al-Qur'dn, MS Medine 179, n6 v -7 r .

PRINCIPLES OF EXEGESIS

151

\ by meditation (tafakkur/ta'ammul/nazr), mutashabihdt by research {talab\
bahth); muhkamdt could be known, mutashabihdt not at all. Finally,

thought to eliminate the contradictions inherent in the preceding series.
For MaturidT, it may be noted, the antithesis zdhir : bdtin meant not the
'literal' as opposed to the 'concealed' significance of the verse in question,
but rather the 'apparent' as contrasted with the 'real' meaning, a dis-
tinction which became the point of departure for his methodological
application of ta'zoil. 1

It was the explicit relating of mutashabihdt to muhkamdt, the latter in
the role of exegetical point d'appui, that provided a foundation for both
halakhic and masoretic exegesis. Jassas (d. 370/981) considered that rela-
tion central to the task of exegesis, but admitted that not every possible
meaning (ma'nd) or aspect (wajh) of the mutashabihdt could thus be dis-
covered: JaiUl Aj)!\ clt> J j^foJI AjUUJL jf >i-U c~ii
oUjtv J-j-j p5icJ1 (Jl eij <— *^=i iS'^\ (J-*2*JI. 2 The procedure
of referring mutashabihdt to muhkamdt might entail reasoning ('aql) or
recourse to authority (sam'), though the former could not be the un-
disciplined application of independent reason, but rather the rational
employment of scholarly tradition. 3 Juxtaposition of muhkamdt and muta-
shabihdt involved explicit recognition of analogy as an exegetical principle,
whether textual or doctrinal, and this became the cornerstone of scriptural
interpretation. 4 That the initial impetus in the application of analogical
deduction to scripture was halakhic, rather than masoretic, might be
thought corroborated by the ascription to Shafi'i (d. 204/820) of the earliest
work entitled Ahkdm al-Qur'dn. 5 More systematically formulated even
than the work of Jassas was the Ahkdm al-Qur'dn of Ibn 'Arab! (d. 543/
1 148) who, in his commentary to the Sahih of Tirmidhi reiterated the
relation of analogy obtaining between muhkamdt and mutashabihdt and
drew attention to the spiritual value of the exegetical activity which must
result from inclusion in scripture of the two kinds of verse. 6 On the neces-
sity of that differentiation Ibn Qutayba had been even more explicit: jjj

Jo\j?ei\ cJUj az*uJ\ cJslmj ,^LJI ^ j^UeJI.*

1 See below, pp. 154-8. 2 GAS i, 444-5 ; Ahkdm al-Qur'dn ii, 3.

3 Ahkdm ii, 5-6. * See below, pp. 166-70.

5 GAS i, 484-90, esp. 480-90 no. VII; the recension is that of BayhaqI (d. 458/1066),
from which it may be necessary to conclude that only the organizing principle, not its
application/illustration, can be dated as early as the end of the second/eighth century.

6 GAL i, 412-13, Suppl. I, 632, 663, 732 ; on the margin of Tirmidhi, Sahih xi, 1 14-20.

7 Ta'teil, 62.

153

QURANIC STUDIES

This view of scriptural exegesis as a divinely imposed task inherent in
the very structure of the document of revelation exhibited considerable
advance towards the scientific formulation of interpretative method and
away from the haggadic division of God's word into prescription, narrative,
and paraenesis. The obligation to study, to immerse oneself in religious
science (rasikhfil-'ilm), could be and was derived from the much-disputed
segmentation of Q. 3: 7, namely, whether interpretation {ta'vnl) of the
mutashabihat was limited to God alone or to God and to those firmly
rooted in religious knowledge. Ibn Qutayba argued that since the prophet
must have known the meaning of those verses ( !), such was necessarily
transmitted to his companions and thus made accessible to the community. 1
The question of juncture (isti'naflibtida") in Q. 3 : 7, between alldh and
wal-rdsikhun, may be understood to symbolize all argument about the
limits of exegetical activity. Insistence upon a disjunctive syntactical value
for the particle waw, as articulated- by SuyutI, 2 was neutralized by the
admission that not every facet of their manifold significance could anyway
be wrung from the mutashabihat. 3 That something of the mystery of
revelation should be reserved to its author could be accepted without

encroaching unduly upon the domain of the exegetes : <)oULxJI ^ja*> ol (Jlj

(JUi' 4^1 b Ja^t*. 4 Their unceasing effort to understand was not thereby

circumscribed, concisely expressed by Zamakhshari a d Q. 3 : j. s That the
ultimately prevailing point of view should be identical with the attitude of
Rabbinic Judaism towards the study of scripture will, in view of all that
has so far been adduced, hardly surprise. Authority was after ail provided
by scripture itself, e.g. Ezra 7: 10, Nehemiah 8: 7, where the basic in-
gredients of a technical vocabulary were also found. 6 Isolated attempts,
already remarked, in Muslim exegetical literature to equate muhkam with
zahir (obvious, in the sense of unambiguous : ma' no. wahid) represent a
polemical tendency, if not specific school disputes, and may be compared
with such disarming statements as DTK •>» pV?D mw mat or JOpa pK
, T , » NXT 1 . 7 It is at least not beyond reasonable doubt whether the
terms peshatjpeshut signified the 'obvious' or 'literal' meaning of scripture,
though in the context of dispute they might tendentiously be so used:
more realistic are Loewe's proposals 'familiar' in the sense of widespread,
and 'authoritative' in the sense of interpretation sanetified by tradition. 8
Save in the polarity zahir :batin employed for allegorical exegesis, Muslim

1 Ta'vnl, 72-4. 2 Itqdn iii, 5-6, also i, 253.

3 e.g. Jassas, Ahkdm al-Quran ii, 3.

4 Suyiiti, Itqdn iii, 9.

s Kashshdfi, 337-8; translated in Goldziher, Richtungen, 127-9.

6 See Zunz, Vortrage, 13-36; Elbogen, Gottesdienst, 194-8.

7 Bacher, Termmologie i, 98, ii, 103, 173.

8 'The "plain" meaning', 158-9, 167, 176-82; cf. Bacher, op. cit. ii. 172-3.

PRINCIPLES OF EXEGESIS

153

use of the term zahir signified 'obvious' only in the sense that one's own
argument was felt to be more compelling than that of one's adversary. 1

Identification in Q. 3 : 7 of the muhkamat with umm al-kitab was uni-
formly understood to refer to the divine archetype of scripture, i.e. its
nucleus (asl al-kitab), analogically deduced from the two other Quranic

occurrences of the phrase umm al-kitab: 13: 39 i__>t53l j»t . . . <U)I

and 43: 4 L>Jd <»jbS0l ^] J aJlj. Though it may be objected that the

deduction was facile and hardly substantiated by the respective contexts of
the locution, the only modification ever proposed was that the muhkamat,
containing as they did divine prescription, enjoyed priority of rank over
verses which were not regulative, and in that sense could be designated
'mother of the book', a phrase often and for quite different reasons applied
to Surat al-Fatiha. 2 Only in Q. 3 : 7, where it could be an interpolation,
may umm al-kitab refer not to scriptural archetype, but rather to exegetical
point d'appui. Such of course would have been more satisfactorily expressed
by a construction *umm lil-kitab exhibited in the Rabbinic precepts
mioaV/Nlpa 1 ? DS V. Horowitz was probably right to reject the equivalence
on the grounds that phraseological similarity was belied by their quite
different applications. 3 On the other hand, that view of the muhkamat as
reference imaradd) for interpretation of the mutashabihat might well be
thought comparable to the Rabbinic notion of em as 'authority'. 4
Allusion to the ultimate necessity of exegesis is contained also in Q. 75 :

Hi

19 jJ Lj LIf. (jj jj. The two other Quranic occurrences of bayan, in 3 : 138
and 55: 4, as well as the single instance of tibyan in 16: 89, designate
'sign', revealed as guidance and mercy. 5 Such also is the function of the
exclusively substantival bayyinajbayyinat, while the participle mubin is in
scripture employed only as attribute (as also adjectival bayyin in 18: 15
and bayyina in 2: 211 and 29: 35). In Q. 75: 19 bayan was understood by
exegetes to signify not merely 'clarity' but also 'clarification', that is,
equivalent to tabyin. 6 In that verse the agency was divine, as in all occur-
rences of the transitive finite forms, which have as object the word 'signs'
(aydt ; e.g. Q. 24: 18, 58, 59, 61) or a noun clause (e.g. 16: 44, 64). Bayan
as exegesis was thus sanctioned by scriptural usage, though the Quranic
locution had not quite the paideutic sense of Biblical hebin, e.g. Nehe-
miah 8 : 7-9, Daniel 1 1 : 33J Synonymous with Quranic bayan is the term
tafstl, also 'clarification' and restricted to the agency of God: either the

1 See below, pp. 242-3.

2 e.g. MaturidI, Ta'zoilat, MS Medine 179, H7 r ; Suyuti, Itqdn iii, 9-10.

3 Untersuchungen, 65 ; Torczyner's conjecture, apud Augapfel, ' "Kitab" ', 387,
presupposed in any case a misunderstanding.

* See Zunz, Vortrage, 338 n. (b); Bacher, Temanohgie, i, 1 19-21.

5 See above, I pp. 5-6. 6 e.g. Zamakhshari, Kashshdf iv, 66i ad loc

7 Cf. Gertner, 'Terms', 21 n. 3.

154

QURANIC STUDIES

book makes all things clear (Q. 6: 154) or is itself made clear (6: 114) by
the act of revelation. The notion of being made distinct by separation/
demarcation (cf. Q. 11: 1, 41:3) provided a technical term for Quranic
periodization. 1 Interpretation of fassalajtafsil as separation and hence
specification is reminiscent of Rabbinic and [Qumranic peresh: 2 the
Muslim term was employed predominantly in halakhic exegesis.

The generic designation of Quranic exegesis is in scripture itself a hapax

legomenon: Q. 25: 33 jJ^JIj iJbio. J^Uj Sjyli ^j. The

subject of yatunaka is 'those who reject/disbelieve' {alladhina kafarU in the
preceding verse), and the entire passage an assurance that opposition to
God's messenger will be countered by divine assistance. The unique context
of the term tafsir is thus polemic, of a kind frequently alluded to in the exe-
getical tradition. 3 Zamakhshari's gloss taksktf referred to the 'uncovering'
of a (maliciously) concealed truth, and represents a standard charge in
sectarian disputer* A similar, but rather more academic, lexical exercise is
contained in the etymology by metathesis {tafsir :tasfir — 'unveiling') pro-
posed in a commentary to Maturidi's TdvAlat? But it seems more than
doubtful that for the technical term tafsir either the Quranic verse 25 : 33
or the metathesis exhibits an authentic Sitz im Leben. 6 It may further be
doubted whether tafsir ever meant, or could really mean, uncovering, in
the sense of bringing to light a concealed significance. The hermeneutical
process involved in tafsir becomes clearer from examination of what
became the standard binary opposition tafsir :ta' mil. There the several
attempts to define the contrast were based on primarily epistemological
considerations. In a synthesis of pertinent arguments Suyuti established a
dichotomy of exegetical modes in which tafsir was defined as the trans-
mission of authoritative witness, scil. to the occasions of revelation (riwayaj
santa/skahdda), and ta'zvil as the product of research and expertise, scil.
in the analysis of scripture (dirdya/istinbat). 7 The polarity had found dia-
grammatic expression in the work of Maturidi: tafsir belongs to the com-
panions of the prophet, ta'wil to those learned in doctrine (ijlpcyiM j^JciS
clgjiill Jjjtdlj). 8 Now, it is hardly possible that these comparatively formal

1 See above, III pp. 116-1 7 ; fdsilajfawdsil.

2 e.g. Kashshdf IV, 184 ad Q. 41 : 3 (with formal but unnecessary reference to a reading
without tashdid); cf. Gertner, 'Pharisaioi', 254-5.

3 See above, pp. 133-7; and I, pp. 36-8.

* Kashshdf Hi, 279 ad loc. ; for kitmdnltahrif/tabdil, see below, pp. 189-90.

5 Samarqandi (ca. 540/1145); see Gotz, 'Maturidi', 35-6; adduced anonymously by
Suyuti, Itqdn iv, 167.

6 See below, pp. 233-5; Wieder's relating Karaite (?) mashpirim to Arabic tasfir might
be linguistically sound, but it may be noted that the Arabic word was never an exegetical
term, and further, that in Arabic lexicology this kind of etymology (metathesis) need not
be taken seriously, see Wieder, Scrolls, 59 n. 4.

7 Itqdn, nam' 77: iv, 167-73; Kitdb al-Mabdnl, ch. VII, 172-82.

8 Ta'wildt, MS Medine 180, i v ; cf. Gotz, 'Maturidi', 31-8.

PRINCIPLES OF EXEGESIS

definitions antedate by much the generation of Tabari and Maturldi (end
of the third/ninth century). But the antithesis tafsir :ta'wil appears in
inverted form in a rudimentary classification of exegesis at the beginning of
Muqatil's Tafsir where, on the authority of Ibn 'Abbas, it was stated that
tafsir is what was known by scholars ('ulamd') and ta'uiil by God alone. 1
To the problems attending the recension of Muqatil's Tafsir, it may be
added that the same tradition from Ibn 'Abbas was adduced by Suyuti,
but contains the term tafsir throughout. 2 It is perhaps not without interest
that in his exposition of the tafsir :tawil polarity Suyuti employed as
generic designation of exegesis the terms bayan and i'rab, explaining that
the use of i'rab in the sense of grammatical sign (hukm nahwt) was in fact
a neologism. 3 That i'rab could, in the light of philologists' use of a'rab
and of Quranic 'arabi, signify 'clarification' is not at all unreasonable,
and that usage was embodied in the titles of the commentaries ascribed
to Farra' and Zajjaj. 4

In addition to the almost purely formal criteria represented by the
riwaya : diraya contrast, a substantial distinction between tafsir and tdvnl
was also formulated. Tafsir was methodologically limited to scriptural
passages bearing but a single interpretation, ta'wil to those bearing more

than one: IS (JjjUllj (sic) IS ^...arlU. 5 The operative terms

were zoajhjvmjuh (aspect) and rajihjmarjiih (prevailing/preferred), used
respectively to designate the range of options and of those the optimum.
As with most technical terminology ultimately associated with a particular
discipline, the locution dhu wujuh retained its earlier and general
significance, and could refer simply to the many facets of the Quranic

message, e.g. ji JjJS jTyJI or ji <JC oTyUI. 6 The proximity

of vmjuh to the Tannaitic panim was pointed out by Goldziher, in respect
of which it may be noted that both terms were employed in halakhic as
in other types ot exegesis. 7 It could be argued that the distinction be-
tween tafsir and ta'tml remained a theoretical one; Abu 'Ubayd, whose
interest in the text of scripture was primarily halakhic, had asserted that

1 Tafsir, MS H. Husnu 17, 2'.

% See above, pp. 1 43-4 : the introduction is full of technical terms which seldom or never
appear in the body of the work ; Itqdn iv, 188: i.e., some tafsir can be known to men, other
tafsir only to God.

3 Itqdn iv, 172-3; istildh hadith; see above, III pp. 109-11.

4 See above, HI pp. 93-4, 98-9; e.g. MS Nurosmaniye 459 and 113, respectively; cf.
GAS i, 48-9.

5 MaturidI, Ta'mldt, MS Medine 180, i v ; Suyuti, Itqdn iv, 167.

6 Suyuti, Itqdn iv, 184; Nahj al-balagha, cited Goldziher, Vorlesungen, 41, 74, n. 4;
cf. also Abbott, SALP ii, 100 n. 48.

7 Richtungen, 84-5; cf. Bacher, Terminologie ii, 157, and on panim as synonym of
te'amim, i, 15 1 ; for the locution md 'l-viajk, as caique of mak ha-ta'am, an example may be
found in Zamakhshari, Kashshdf i, 519-20 ad Q. 4: 48.

156

QURANIC STUDIES

they were one and the same. 1 The difficulty lay in determining which
Quranic verses might be characterized as containing more than one
aspect (wajh) and hence suceptible of interpretation by ta'wil. From
Q. 3 : 7 it was clear that ta'wil was applicable only to the mutashabihat,
identification of which was, as noted, remarkably unstable. The analogical
relationship seen by Maturidi and Jassas to obtain between muhkam
and mutashabih meant that in practice the latter could be explained by
reference to the former, even though not every aspect of the muta-
shabihat could be so illuminated. The methodological difference between
tafsir and ta'wil might seem thus at least blurred, if not entirely effaced,
by the admitted interdependence of muhkamat and mutashabihat. The
formal difference could, however, be maintained: in contrast to ta'wil,
which involved investigation and research, tafsir depended upon tradition.
The same solution to a scriptural problem might, in other words, be
reached by different methods: its acceptability was often no more than
a matter of presentation, that is, with or without the requisite witness
(skahadafriwaya). It seems clear that the tafsir : ta'wil dichotomy symbolized
a dispute rather more fundamental than one merely of method or termino-
logy, namely, the exegetical relationship between canonical and non-
canonical material in the witness to revelation preserved and transmitted
by the Muslim community.

The necessity of resort to tradition as interpretative complement to
scripture was the crux of sectarian dispute between and within the Jewish
and Christian communities. The extent to which tradition could be re-
garded as dispensable depended upon the successful elaboration of
exegetical techniques which might be seen to elucidate scripture, as it
were, from within. Without stressing unduly the essential futility of steps
taken to that end, it is worth observing that there is about them a consider-
able measure of uniformity. The 'Torah-centricity' of such groups as the
Qumran sectaries, the Apostolic Christians, and the Karaites generated a
series of interpretative principles which might have been, and in some
instances actually were, freely exchanged. 2 One such common element was
the division of scripture into 'manifest' and 'concealed' parts, the latter
epithet employed not in the sense of esoteric but of ambiguous or equivocal,
in short, deutungsbedurftig. It would, in my opinion, not be unjustified to
see in the antithesis muhkamat : mutashabihat a reflex of niglot:nistarot, and
in particular of the contention that in each pair the second element might
be elucidated by reference to the first. 3

In Muslim practice that process of analogical deduction presupposed

1 Apud SuyutI, Itqdn iv, 167.

2 See Gerhardsson, Memory, 172-3, 333, 284-7; Wieder, Scrolls, 53-62; Rabin, Qum-
ran, 95-" 1 •

3 Wieder, Scrolls, 76-7; Rabin, Qumran, 99.

PRINCIPLES OF EXEGESIS

1 57

the self-sufficiency of the Qur'an, and as such must be regarded as pole-
mical in character. The dispute was articulated, if never quite satisfactorily
resolved, in the elaboration of halakhic exegesis. 1 The notion of ambiguity
in the term mutashabih, as partaking of more than one semantic aspect, was
reinforced by identification with mushtabih, possibly an allusion to the
parallel passages Q. 6: 99 and 6: 141, or others, e.g. Q. 2: 25 and 2: 118, 2
but more probably engendered by its antithetical relation to muhkam. The
synonymity of mutashabih and mushtabih is explicit in Zamakhshari ad
Q. 3 : 7, and had much earlier become axiomatic for masoretic exegesis. 3
It would be misleading, depite the centrality of Maturidi in its formula-
tion, to suggest that scriptural exegesis which dispensed with tradition was
invariably designated ta'wil. In scripture itself the term occurs seventeen
times and, save for Surat Yiisuf (Q. 12: 6, 21, 36, 37, 44, 45, 100, 101)
where it could only be rendered 'dream-interpretation' (ta'bir al-ruya),
it was consistently glossed 'outcome'/'sequel' {'aqiba), thus lending the
term a distinctly eschatological flavour which accorded nicely with the
haggadic definition of mutashabihat as four of the cryptic Quranic sigla*
It seems to me that this eschatological, or at least prognostic, sense fits
rather better than 'interpretation' the use of tawil in the much-cited verse

of 'Abdallah b. Rawaha: -dyp Jp *5"l±S Li" aLjI" J* ^TUbS

It was partly owing to that application that the term ta'zoil achieved
enduring status in allegorical exegesis, which was largely though not
exclusively sectarian. 6

Function as generic designation of scriptural exegesis devolved thus
upon tafsir, eventually employed for most if not quite all varieties of that
exercise. That its Sttz im Leben was almost certainly the lexicon of profane
rhetoric does not of course exclude influence from other quarters. 7
Whether the literary activity of the haggadists was actually described by its
authors as tafsir is, owing to redactional complications, not easily answered.
In his historical survey of exegetical method Goldziher described that
'primitive' interpretation as the nucleus of what became traditionist tafsir
(tafsir manqul) and as secondary to establishment of the text of scripture
( Textgestaltung). 8 That view of the anterior existence of a ne varietur text
gradually subjected to interpretation reflects of course the 'Uthmanic
recension traditions with all their very familiar implications. A comple-
mentary feature of this traditional view of Islamic origins is the assertion
that early attempts to interpret the text of scripture were frustrated by an

1 See below, pp. 175-7, 188, 201-2. 2 Cf. Ibn Qutayba, Ta'toil, 74.

3 See below, pp. 212-15; Kaskshdfi, 337-8. 4 See above, p. 149.

5 Tabari, Annates I/1595; cf. Goldziher, Studien ii, 112 n. 5; id. Abhandlungen i,
60 n. 2; id. Richtungen, 278.

6 See below, pp. 243-6. ' See below, pp. 233-5.
8 Richtungen, 55-85 and 1-54, respectively.

i5«

QURANIC STUDIES

official prohibition, or at least restriction, of exegetical activity. 1 Reason
for the official measures, invariably associated with the figure of 'Umar b.
Khattab, was seen to be an expression of extreme piety. Modifications of
this view have subsequently appeared, notably those of Birkeland and
Abbott. 2 Birkeland recognized that the alleged opposition was a late
formulation exhibiting school disputes about the form in which tafsir
ought to be transmitted, a contention often and misleadingly expressed
in terms of the opposition tafsir bil-ilm : tafsir bil-ra'y. 3 Abbott's insist-
ence upon the historicity of the story of 'Umar and Sabigh seems to
me to have missed the point entirely, as do her simplistic references to
mutashabihat and tafsir al-nabi*

The figure of 'Abdallah b. 'Abbas (d. 68/687) as tarjumdn al-Qur'an
might be thought to pose something of a problem. Birkeland's relegation
of that figure to a personification of consensus (ijma), symbolized in
particular by the isnad Ibn Sa'd-Ibn 'Abbas, is a reasonable hypothesis
indeed, especially if qualified by an admission that the historical process
reflected in tafsir cannot be reconstructed before the beginning of the
third/ninth century. 5 References in the tabaqat literature to earlier author-
ities, almost without exception disciples of Ibn 'Abbas transmitting on his
authority, can hardly be said to represent more than the proliferation of
companion isnads shown to be characteristic of legal traditions. Impetus
for the production of both legal and tafsir traditions was halakhic, and
objections to 'tafsir' are in my opinion to be understood only secondarily
as disapproval of independent reasoning (ra'y) as opposed to traditional
science ('ilm). The primary dispute was about the sources of doctrine
(usul al-fiqh) and reflected in the respective claims to priority put forward
by advocates, on the one hand, of canonical revelation, and on the other,
of non-canonical revelation. 6 The role of 'Umar in the anti-tafsir traditions
might be compared to that of 'Uthman in the canonization traditions : an
explanatory mechanism designed to attest the earliest possible origins for
the components of Islam. 7 The absence of scriptural interpretation before
the generation of haggadists does not require an explanation so contrived
and internally inconsistent.

That during the second/eighth century halakhic disputes were essentially
ones about principles as opposed to methods may be thought corroborated

' Richtungen, 55-64. 2 Muslim Opposition; and 5 ALP ii, 106-13.

3 Muslim Opposition, 28-32.

4 SALP ii, 106-13; that the story of 'Umar and Sabigh constituted proverbial illustra-
tion of tiresome interrogation, not only about the mutashabihat, but also halakhic material
in scripture, seems clear from Malik's reference in a discussion of the spoils of war,
Muwatta', 455: Kitdb al-Jihdd no. 19.

5 Muslim Opposition, 32-42; I am not quite certain, however, whether Birkeland would
accept that qualification.

6 See above, I pp. 51-2; and below, pp. 161-3, 188.

7 See BSOAS xrri (1968) 613-16.

PRINCIPLES OF EXEGESIS

• by the witness of two documents often adduced as milestones in the
juridical and political development of Islam. The first of these is
the Risala fil-sahaba of Ibn Muqaffa' (d. 142/759). 1 The standpoint of the
author in matters pertaining to the practical administration of justice was
characterized by Schacht as a recommendation of procedural uniformity,
to be imposed by the government (sic) upon a situation of juridical chaos. 2
The elements of confusion could be identified as regional dispute, about
the priority of sunna (as ius consuetudinis) and ra*y (as practical inference),
overlaid by the legacy of Umayyad administrative practice. It may well be
that Ibn Muqaffa"s proposals were derived from Persian models, but the
notion of charismatic leadership underlying his emphasis upon the
position of the caliph hardly required foreign inspiration. 3 My interest
here lies exclusively in the role of scripture in Ibn Muqaffa"s suggestions
for organizing the Islamic community. As a source of caliphal authority
the Qur'an (designated kitab) received but scant attention, usually in tan-
dem: al-kitab tmlsunnaf and was only once cited, as 'revelation' (tanzil),

i.e. Q. 7: 43 jUII UIja jf VjJ <S-^-&^ L>_? m a context which re-
commended, appropriately, recourse to sound reasoning. 4 Arbitrary em-
ployment of reason was condemned, but also arguments based upon sunna
which could not be derived from the prophet or from one of his rightly

guided successors (!) «Juu ^ <£JigJI iJl jl <UJl x§z ^c. s Sound

reasoning consisted in the caliphal application of analogy (tadbir zoa-qiyds),
and in matters of dispute the caliph was to employ that instrument: JixJ
J.X*Jli (jJ^Sl <4-ifj (Ji-^Jl; ij' -M- 6 A reference to the

significance of ahl al-fiqh wal-sunna wal-siyar wal-nasiha leaves a distinct
impression that the function of scholarship was to enhance the caliphal
authority. 7

Now, the tenor of this official communication might be thought to
corroborate a good deal of similar evidence that at mid second/eighth cen-
tury revelation had yet to achieve status as recognized authority for doc-
trine or for policy within the Muslim community. 8 The appeal to analogical
reasoning reflects a stage of doctrinal development prior to the onset of
formalized transmission of authoritative opinion, and a period in which
the celebrated instructions to the qddi of 'Umar b. Khattab might well have
been composed. 9 Further evidence of Ibn Muqaffa"s attitude to Muslim

1 GALi, 138, 151, Suppl. I, 210, 236; in M. Kurd "All, Rasd'il al-bulaghd', 117-34.
a Origins, 58-9, 95, 102-3, *37-

3 Schacht, Origins, 95; cf. Goitein, 'Turning point', Studies, 149-67, esp. 163-4, where
the theory of a Persian model is modified.

4 Risala JU-SaJmba, 121, 123, 122. 5 Risala, 126.
6 Risala, 127. 7 Risala 127.
8 See above, I p. 44. 9 See above, II pp. 56-7; and below, pp. 166-7.

i6o

QURANIC STUDIES

scripture might possibly be elicited from the refutation of arguments as-
cribed to him by the Zaydl imam Qasim b. Ibrahim (d. 246/860). 1 Attribution
of the work being refuted was rightly considered by Guidi to be very ques-
tionable indeed. 2 So little of that ascribed to its putative author is preserved
in the refutation that it is quite impossible to say more of Qasim's adversary
than had he not existed he would have had to be invented. References to
the document of revelation consist exclusively of ingenuously literal inter-
pretations of Quranic phraseology, e.g. descriptions of God as wager of
war, as destroyer of the utnam khdliya, as seated upon His throne, etc., 3
each deftly and in turn thrown out of court by Qasim, whose display of
expertise in the metaphorical exegesis of anthropomorphic expression was
faultless. That the author of the Risala fil-sahdba, a work both formally
pious and substantially sophisticated, could have been responsible for the
trivial argumentation attributed to him by Qasim is unlikely. At several
points the latter found occasion to criticize severely his opponent's know-
ledge of Arabic, and accused him of having composed a barbarous book
{a' jam al-baydn)A Childishly inept humour like the alleged beginning of
the book : 'In the name of the merciful and beneficent light' {bismi 'l-nuri
'l-rahmdni 'l-rakim) can be taken seriously only as caricature and point,
not so much to a mu'arada of the Qur'an (traditionally ascribed to Ibn
Muqaffa'), as to an ideal target for the vituperative criticism of the
Mu'tazili Qasim. 5 The major portion of the refutation, thus presumably
also of its fons et origo, consists of polemic about the principles of divine
creation and of God's justice and retribution, suspiciously appropriate, it
might seem, to a Mu'tazili tour deforce.

Significant in a way quite different from the Risala fil-sahdba of Ibn
Muqaffa' is the work entitled Risala fil-qadar and ascribed to Hasan
Basri (d. no/728). 6 Its authenticity and ascription were accepted by
Ritter and by Obermann; and its authenticity, if not ascription, by
Schacht. 7 The argument of the tract, essentially dogmatic and in fact little
more than identification of Satan as agent and repository of Evil, was found
offensive by Shahrastani (d. 538/1143), who was thus willing to ascribe it
to Wasil b. 'Ata' but not to Hasan. 8 Schacht remarked the exclusive em-
ployment of Quranic shawahid in the work and a concomitant absence of

1 GAS i, 561-3 ; Kitdb al-radd 'aid 'l-zindlq, in Guidi, La Lotta, arabic pagination 3-55.

2 La Lotta, viii-xi, though he judged it to be a typical product of the period in which Ibn
Muqaffa* lived, cf. xxi-xxiii; also, Nyberg, 'Zum KampP, 425-41, esp. 431-2.

3 La Lotta, arabic pagination 17-26, 29-31, 35.

4 La Lotta, arabic pagination 8, 10, 31, 33, 39-40, 43.

5 La Lotta, arabic pagination 8; cf. Goldziher, Studien ii, 401.

6 GAS i, 591-4, esp. 592 no. 3; MS Koprulii 1589, Ayasofya 3998; in Ritter, 'From-
migkeit', 67-82; for an analysis of its contents, see Schwarz, 'Letter', 15-30.

7 Ritter, 'Frommigkeit', 62-4; Obermann, 'Political theology', 138-62, esp. 154-8;
Schacht, Origins, 74, 141, 229.

8 Kitdb al-nulal wal-nihal, on the margin of Ibn Hazm, Kitdb al-fisal i, 59.

PRINCIPLES OF EXEGESIS

161

traditions from companions or prophet, arguing e silentio for composition
at the beginning of the second/eighth century, chronologically consonant
with the scholarly activities of Hasan. 1 There are, indeed, ninety-five
verses and parts of verses adduced (five of them twice), from thirty-nine
well-spaced suras of the canonical text of revelation, a few passages con-
sisting of nothing more than a concatenation of scriptural loci. 2 None of the
scriptural material may be described as regulative, but rather, as admoni-
tory and paraenetic, drawn from eschatological contexts stressing the
ethical implications of the Quranic theodicy. The over-all structure of the
risala is polemical rather than halakhic, formulated as epistolary address
and regularly punctuated by imperatives beginning 'O commander of the
faithful', with one passage based on sustained employment of the apo-
strophic 'And if you were to say . . .' (wa-law qulta). 3 The main body of the
treatise contains a series of allegedly disputed points in scripture, thirteen
in all and introduced 'There is dispute about His word' (fa-yujddilunj
yunazVun fi qawUhi), in which the author disposes in a tidy though facile
manner of his anonymous adversaries. 4 Those disputes could have no basis
in fact, since scripture (kitdb allah generally; qur'an occasionally) contained
neither inconsistency nor contradiction unless it had been tampered with
(wa-harrafiihu). s Reference to the reliability and omniscience of scripture
is sufficiently recurrent and emphatic as to provoke the question of the
author's real purpose. It might seem that the very absence of all but
scriptural shaivahid and the express insistence that all answers were to be
found therein could suggest an usul controversy, in which the 'plain mean-
ing' of scripture was being asserted in the face of analogical reasoning and
of tradition, whether from companions or prophet.

Some support for this conjecture is found in what may be designated
the 'framework' of the risala. The request of the caliph 'Abdalmalik for
information on the problem of qadar (liberum arbitriutn) was constructed
round his wish to know whether Hasan's knowledge had been derived from
tradition(s) from companions of the prophet, his own opinion/reasoning, or

from an argument confirmed in scripture: <_>Uv>t fr* ^ i>ljj
jTyU! (j AiJiyoJ (J^o jj tf. -UjIj tsfj *AJl (Jj-.j. 6 The reply

was interestingly circumstantial: he (Hasan) had learned from his pre-
decessors who lived according to the word of God and transmitted His
wisdom, who followed the surma of the prophet (sic), who knew right
from wrong, and who did not assert other than that which God had

Himself expressed for the benefit of mankind in His book: ^1 U b5~js]

\jZU.\j <Ca£». IjJjJ Aj}\ jAs Ijiufr (jjjJI tJuJI ^* <U)I di?sL*»t £-JU£j|

1 Origins, 229. 2 Ritter, 'Frommigkeit', 71, 73, 82. 3 Ritter, 75.

* Ritter, 72-80. 5 Ritter, 70 line 15-71 line 2. 6 Ritter, 67 lines 8-9.

4SS9C75 G

QURANIC STUDIES

Similarly, in the 'covering letter' from Hajjaj to the caliph recommending
Hasan's treatise, the Quranic sources of Hasan's views were stressed by
means of a paraphrastic conflation of the caliph's own inquiry. 2 Thus,
prophetical Sunna and companion traditions were included there as

respectable sources of knowledge, though inferior to scripture: oyoj

JU.I <Jj| ^ifrlj «U)I <J>«j tjlxT' j «*jjjOv=J

^3 (j Ajailj ,JU.J <UJb ^iil y> JbJ «UJ| Jj-j <_>b**>l jj-j ^Ul u iLJ I
{ y~*l\ ^ <Ull t->^ IX* Protestations like 'Any argument not
based on scriptural proof is fallacious' (^IxT^ ^ ji Jy
'ti'fe «U)I) 3 or 'Thus does scripture speak, O commander of the faith-
ful' ((jkis aM\ <~>\^T £~u|JI ^-.f U Ioa) 4 might seem to indicate a more

than merely casual concern for usul priorities. Reference to the defor-
mation (tahrif) of scripture is not infrequent, 5 as also to arbitrary or
otherwise unsatisfactory exegesis, 6 twice specified as interpretation by

personal opinion/reasoning: dUi OjJjW and p&J j tLU3 OjjZiJ. 1

That particular charge can hardly be taken seriously, since it was the
method employed by the author himself throughout the treatise, and the
epithet may have been nothing more than a tag for opinions from which
he dissented. His own exegesis consisted in the ingenuous assertion of
scripture's 'plain meaning', easily apprehended by the unprejudiced eye
and a basic knowledge of Arabic. Though quite unconcerned with
linguistic or rhetorical analysis, he adduced in two instances examples

from the usus loquendi (kalam al-'arab): ad Q. 19: 59 jjj Ljf Uj^
UJl Uj*.

and 3: 178, where a line of poetry was offered and the locution quran
'arabi apostrophically (and predictably) explained as a concomitant of Arabic

1 Ritter, 68 lines 3-6. 2 Ritter, 80-1 : note (MS Ayasofya 3998).

3 Ritter, 68 lines 13-14; cf. Schacht, Origins, 141.

* Ritter, 69 line 13 ; curiously, the parallel construction in Q. 45 : 29 was not adduced by
the author of the risdla, i.e. hadhd kitabuna yantiq 'alaykum bil-haqq; while in the risala
reference to scripture is unmistakable, it could be of interest to note that Qummi, Tafsir
ii, 295, emended kitabuna to bi-kitdbina, thus making the prophet, not the book, the spokes-
man of God, probably a reflex of Imami apologetics, similar to the emendation
umnta: a'imma in Q. 2: 143 and 3: no, Tafsir i, 63 and no, respectively; cf. Goldziher,
Richtungen, 281-2.

5 Ritter, 68 line 11, 69 line 19, 70 line 16; see below, pp. 189-90.

6 Ritter, 69 lines 19-20, 75 line 2 (ta'wil); 74 line 2, 78 line 16 (ta'awwul).

7 Ritter, 74 line 6, 78 line 5 : taavmala and fassara synonymous.

8 Ritter, 79 lines 8-9.

PRINCIPLES OF EXEGESIS 163
. eloquence p^-kU. i^jt. Jl -Mil <Uyl l ^JUjJI j~J b jTyiJIj

Method and style in the risala, unblemished by textual and semantic
problems or scholarly apparatus, are not unlike those of the haggadists,
characterized by straightforward equations and absence of authorities.
Occasional employment of the connectives ay and ya'ni, even yaqUl, 2 may
be thought to attest to that similarity, corroborated by a simplistic inter-
pretation of the Khidr verses (Q. 18: 60-82) in order to demonstrate the
compatibility of free will with divine foreknowledge. 3 In the light of
explicit reference to companion tradition, prophetical Sunna, with which
may be contrasted mention of sunnat allah in Q. 40: 85,* and exegetical
ra'y, it could be argued that the exclusive employment of Quranic loci was
neither fortuitous nor a consequence of the early composition of the risala. 5
I am inclined to assign the treatise to the end of the second/eighth century,
after the development of haggadic exegesis and during the period of usul
disputes represented by, at least, the conflict between traditionists and
Mu'tazila. Ascription of this hortatory and edifying work to a figure like
Hasan Basri hardly requires explanation, and may be compared to a similar
tendency serving the reputation of 'Abdallah b. 'Abbas.

One point in the risala deserves further mention. The author's assertion
that the text of scripture was free of inconsistencies and/or contradictions
rested upon two assumptions: first, that such as might be found must be
the result of (malicious) alteration, second, that the meaning of mutashdbih
was 'analogous', in the sense of mutually corroborative (as in Q. 39: 23):

6jjj>j <U)I «wjU5^ js^J^JI b i^iJU. JaIj

bi*j a^ojo vJJUbj Urt-; 6 Thus, the antithesis muhkamimutashabih

could, once formulated, be adduced in justification not only of 'applied'
but also of 'pure' exegesis. 7 Preoccupation with apparently contradictory
statements in scripture generated two separate but interdependent views
within whose terms a very carefully delimited typology of contradiction
(ikhtildf/tanaqud) could exist side by side with what might be called a
recognized set of 'standard puzzles'. What must be the earliest version of
the latter is found in the form of an appendix to Muqatil's Tafsir al-
khamsimi'at aya. 8 The relation of the appendix, contained in folios ioo v to

1 Ritter, 76 line 20-77 hne 3- 2 Ritter, e.g. 79 line 8, 68 line 15, 67 line 12.

3 Ritter, 77 lines 12-19; see above, p. 128. 4 Ritter, 79 lines 3-4.

5 Pace Schacht, Origins, 74, 141, 229; in this connection the observation of Obermann,
'Political theology', 142 n. 11, that these technical terms here make a remarkably early
appearance, gains fresh and certainly unintended significance.

6 Ritter, 70 line 15-71 line 2. 7 See above, p. 148.
8 GAS i, 37 no. 1 ; MS British Museum Or. 6333 ; Abbott, SALP ii, 96 ; cf. BSOAS

xxxi (1968) 614.

164

QURANIC STUDIES

io3 r , to the body of the work is something of a problem. For the copyist
of this manuscript (an unicum, dated 4 Jumada I 792/20 April 1390) the
appendix clearly belonged to the main text (same paper, ink, hand, etc.),
but was separated therefrom by a fresh basmala and two abrupt changes of
subject. The style is haggadic and almost certainly that of Muqatil (or of
the works traditionally ascribed to him), beginning with a ta'yin of Q. 18:

60-82 identical to that of his major tafsir: 'i~ju. JS" A^-tj jfc' <^a)I ^.1
l-LL. ^1 Uai.. 1 Once the dramatis personae of the Musa-Khidr story
were identified, a new isndd introduced the topic of scriptural contradiction
(ikhtildf al-Qur'an), related on the authority of Ibn 'Abbas, and illus-
trated by the following nine problems: 2

1 . Whether on the day of Judgement there shall be communication between those
being tried: Q. 23: 101 vs. 37:27 (37:50, 52:25).

2. Whether on that day polytheists will or can avail themselves of the services
of their deities: Q. 6: 22 vs. 6: 23.

3. Whether God created heaven or earth first: Q. 79 : 27-30 vs. 41 : 9-1 1.

4. Whether the grammatical value of kdna in Q. 4: 23, 4: 134, etc. reflects a
permanent or a temporary quality.

5. Whether the promise in Q. 20: 124 (he who ignores my admonition shall live
in penury) is confirmed or belied by experience.

6. Whether the promise in Q. 16: 97 (goodness shall be rewarded) is confirmed
or belied by experience.

7. Whether on the day of Judgement those being tried shall be asked about their
misdeeds: Q. 55 : 39 vs. 15: 92-3.

8. What is the precise meaning of 'guidance' (hudan) in God's warning to Adam
and Eve: Q. 20: 123 ?

9. What is the precise meaning of 'before him' (amdmahu) in Q. 75 : 5 (but man
desires to sin before him) ?

The solutions proposed by Ibn 'Abbas to these problems were un-
equivocal and unsophisticated, and evoked a primitive level of popular
discourse. 3 The exegetical principle involved in the elimination of apparent
contradictions was that which distinguished between different contexts
{mawatin) despite similar or identical phraseology. Thus, nos. 1, 2, 3, and 7
did not contain 'real' contradictions because the opposing verses referred
to quite different situations, or to different aspects of the same situation.
Nos. 5 and 6 were, on the other hand, allusions to eschatological fulfilment
or, alternatively, virtue as its own reward. To the lexical problems in nos. 4,
8, and 9 solutions were evaded by resort to theology: kdna predicated of
God must signify an eternal attribute; 4 'guidance' was the Qur'an; 'before
him' indicated progression in disobedience (qudumanfil-ma'dn).

1 Tafsir, MS. H. Hiisnii 17, i7i v -2 r ; see above, pp. 128, 135-6.
* Tafsir, MS BM Or. 6333, ioo v -i r .

a Tafsir, ioi'-3 r . ♦ Cf. Reuschel, 'Wa-kana llahu 'alimari, 147-53.

PRINCIPLES OF EXEGESIS

\ A schematic and slightly more sophisticated version of what eventually
became the standard (!) scriptural puzzles was attributed to Muqatil and
included in his Kitab al-tanbih wal-radd by Abu Husayn Malati (d. 377/
987).' The extracts from Muqatil, adduced by Malati without supporting
reference to Ibn 'Abbas, are two: 2 on Quranic contradictions, and on
semantic and phraseological correlations, of which the first seems to be a
systematic expansion of the material described above from BM Or. 6333.
I have found no other work of Muqatil in which this material appears, or
in this form could appear, except as appendix or as some other variety
of formal intrusion. 3 In Malati's version, 4 the number of scriptural contra-
dictions was increased to twenty-five, including four examples from
BM Or. 6333 (nos. 1, 2, 3, 7). The principles of harmonization remained
the same, but were enhanced by differentiated and normative expression.

Each solution was, for example, introduced by the formula ^ ju&

I jS~ j~*J£ J U.£x50j jj^aiu^j ,jt>^j Uojo 4*3j«j ^^a'j and the

reason for the apparent discord specified as different contexts (mawdtin
mukhtalifa), similar phraseology in analogous circumstances (sildt al-kalam
mushtabihajwujuh al-halat mushtabihajikhtildf al-halat mushtabih), tempor-
ally separate aspects of the same situation (ttmjiih taqdim al-kalam mushta-
biha), etc. That in exegetical usage mushtabih and mutashdbih were
functionally synonymous has been noted. 5

The manner in which this kind of problem, together with its attendant
loci probantes, became a constant in the literature of scriptural exegesis may
be elicited from Suyuti's chapter on the subject in his Itqdn, where the
notion of 'contradiction' was appropriately described as irresponsible

fantasy and beneath the dignity of God's word ^jUJI ^a^j L il^Jlj
cLUi ,JuJ aS^S'j oUSl ^j, and illustrated by four standard ex-

amples from Ibn 'Abbas (nos. 1-4 in BM Or. 6333). 6 But for SuyutI these
were merely a point of departure for a survey of elaborate rhetorical tech-
niques designed to prove what had been unequivocally asserted, namely,
that the text of revelation contained neither contradiction nor inconsistency.
So stated, that position was manifestly indefensible, in tacit recognition of
which recourse was had to more sophisticated terminology. Ibn Qutayba was
able, for example, to distinguish between contradiction proper (ikhtildf
tadadd) not found in the Qur'an save in cases of regulative abrogation,

1 GAS i, 607; in Dedering, Bibliotheca Islamica IX.

2 Kitab al-Tanbih, 44-56 and 56-63, respectively; see below, p. 210.

3 Pace Abbott, SALP ii, 96; manuscripts have so far been discovered for three, not
four, separate works of Muqatil, cf. GAS i, 37; the extracts from Malati were published
separately, and earlier, by Massignon, Recueil, 194-210.

4 Kitab al-Tanbih, 44-56. 5 See above, p. 157.
6 Itqdn, now' 48 : iii, f^-Sg.

QURANIC STUDIES

and difference by variation (ikhtilaf taghayur) of which an example was the
word umma in Q. 12: 45, expressing both period (hin) and forgetting
(nisyan). 1 This argument did not, however, prevent his adducing a sim-
plistic harmonization of the by that time traditional instance of contra-
diction: in the question of conversation on the day of Judgement. 1 For
Ibn Taymiyya the distinction lay between contradiction proper (ikhtilaf
tadadd) and variation (here ikhtilaf tanazvwu) of which only the latter
could be found in scripture. 3 For KirmanI the antithesis was expressed as
concepts mutually exclusive (tanaqud) not present in the Qur'an, and
relational difference (ikhtilaf talasum), found in scripture as multiple varia-
tions upon a single theme. 4 That proliferation of technical vocabulary is
witness to an abiding concern with textual consistency. It was one form of
exegetical activity in which no serious scholar ever adduced in support of
his argument the 'plain meaning' of scripture.

The concept of ikhtilaf as variation, and not simply as flagrant contra-
diction, presupposed or at least implied some degree of both textual
integrity and conceptual unity in the document of revelation. The manner
in which those were perceived or, perhaps more accurately, were created,
emerges from examination of the work of halakhists and masoretes. Their
primary, and indispensable, instrument was analogy. Under that general
title was subsumed a number of related but methodologically distinct pro-
cedures. The basic distinctions were two : between deductive and inductive
establishment of an analogical relation, and between halakhic and masoretic
application of the instrument. While a general impression that the Arabic
term qiyas denotes an exegetical procedure both inductive and halakhic is
not altogether unjustified, it is also an over-simplification. An example can
be seen in Bergstrasser's selection of qal •ma-homer as illustrative of qiyas
method: 'da sie nicht wie die anderen die Interpretation eines normativen
Textes, sondern die Gewinnung einer neuen Bestimmung aus einer vorhan-
denen regeln soil'. 5 In fact, qiyas was employed both for extrapolation of
fresh principles from existing premisses and for interpretation, as well as
for establishment, of the scriptural text, though the appearance of fixed
and consistent terminology was admittedly later than the phenomena
themselves. 6

What must be the earliest, or almost the earliest, reference to recourse to
analogy for halakhah occurs in the Risala fil-sahdba of Ibn Muqaffa' : ^kui
jA*Jb (j^jl ^JjA^iiXJLj {jJ>±yA\ ,jpj (J| ju». 7 The locution

1 Ta'ml, 31.

2 Tamil, 47 ad Q. 77: 35 vs. 39: 31, one of several sets of contradictory verses on that
subject, e.g. BM Or. 6333 no. 1, Kitdb al-Tanblh, 44.

3 Apud Suyuti, Itqan iv, 176-7. + Apud SuyutI, Itqan iii, 89.

5 Bergstrasser, 'Anfange', 81.

6 See Tahanawl, Istildhdt, 1189-96: qiyas lughawi and qiyas shar'i.
' Risdla fil-$ahdba, 1 27 ; see above, pp. 158-60.

PRINCIPLES OF EXEGESIS

ashbah al-amrayn bil-'adl can only refer to 'equitable comparison of two
cases', in the sense that their respective merits and demerits were to be
juxtaposed. In that particular context it was the reasoning of the caliph
which was evoked to produce a solution. Possibly contemporary with, but
probably later than that passage is the recommendation ascribed to 'Umar

b. Khattab in his instructions to Abu Musa: ix*. <j\j3 ^^J
dti's Jjo j*S pj (Jlljilj oL-ijI k-i^-lj. 1 That the imperative qis

here signified 'juxtapose' was recognized by Margoliouth, who related
that usage to Talmudic hiqqish. 2 Of similar importance in the recommenda-
tion are the terms ashbah and amthal (probably hendiadys), referring to
those common elements in (two) propositions which may be juxtaposed
(scil. for the purpose of comparison). The explicit condition 'in those matters
for which there is neither qur'an nor sunna' suggests, in my opinion at
least, a date later than the corresponding passage in Ibn Muqaffa', which
reads 'in halakhic dispute (ikhtildf al-ahham) either about a matter trans-
mitted from the ancients (ma'thur 'an al-salaf) ... or about a case of arbi-
trary reasoning (ra'y ajrdhu ahluhu 'aid y l-qiyds)'. 3 The use of shabbaha for
analogical juxtaposition is attested elsewhere, e.g. in Malik 4 and in Bukhari. 5
Whatever the linguistic relation of Arabic qas to Hebrew hiqqish, it may
seem that shabbaha was equally appropriate as notional equivalent to the
Hebrew term.

In the contexts adduced above the terms ashbahjashbahj shabbaha can
hardly refer to juxtaposition based on identical or even similar phraseology.
Introduction of a tertium comparationis is at least implicit in Bukhara's
discussion of the legatee's obligation to fulfil a pilgrimage vow derived
from a duty to pay the outstanding debts of the deceased. 6 Employment of
Talmudic hiqqish was characterized by the same latitude: some but by no
means all of the examples assembled by Bacher depend upon a recurring
locution or even textual juxtaposition in scripture. 7 Methodological distinc-
tion in reasoning by analogy with and without a third term became
evident only with the refinement of technical terminology, of which a
valuable illustration is afforded by the evolution of Talmudic gezerah
shawah. That that term came to, but did not originally or consistently,
designate analogy based on occurrence of the same word has been often
and convincingly demonstrated. 8 In halakhic argument the principle of

1 Margoliouth, 'Omar's instructions', 309, 320; Goldziher, Zahiriten, 9.

2 Margoliouth, loc. cit. ; see also Schacht, Origins, 99.

3 Risdla fiUSahdba, 127.

4 Mudawwana ii, 94, cited Schacht, Origins, 117: *to assimilate'.

5 $ahih, 'Kitab al-I'tisam', no. 12, cited Goldziher, Zahiriten, 107.

6 Goldziher, loc. cit. 7 Terminologie i, 44-6, ii, 57-8.
8 See Bacher, op. cit., i, 13-16; Gertner, 'Terms', 24-5 ; Loewe, 'The "plain" meaning',

164-5; and cf. the Pauline application in Romans 4: 3-9, which did depend upon an
identical word, Gerhardsson, Memory, 287-8.

1 68

QURANIC STUDIES

inductive analogy, by which I mean those varieties involving a tertium
comparationis or ratio, is likely to have antedated insistence that the ana-
logue be an identical phrase. It must, on the other hand, be admitted that
contexts held to be analogous could in the first instance have been so
related by reference to common phraseology. The complementary prin-
ciple of deductive analogy may be thought to have had its origin in maso-
retic rather than halakhic argument. 1 In Muslim juridical literature the
appearance of argument based on inductive analogy preceded by about
a century systematic use of the technical term 'ilia for ratio, a notion for
which Shafi'i employed asl (root/base/basis) and ma'na (meaning/sense). 2
With the development of (fundamentalist) opposition to inductive analogy
(called ta'lil), differentiation crystallized in the antithesis ta'lil:mansus, the
latter term being employed, at least by sectaries of the Zahiri school, for
analogy based on a textual similarity. 3 The generic term itself for analogy,
qiyas, came increasingly to be modified by such epithets as were found
necessary to describe the relation to one another of its components, e.g.
jaliyjzahir (explicit), khafiy (implicit), and even by phrases which effaced the
opposition ta'lil:manstts, such as 'ilia mansusa and qiyas ma'qul al-nass*
Much of that proliferation of technical terms reflected (often necessary)
steps to circumscribe the range of analogical argument, frequently un-
disciplined and far-fetched. 5

Masoretic analogy was, on the other hand and by its very nature, mostly
deductive. Though designated qiyas and defined in terms identical to those

employed for its halakhic counterpart, e.g. by Ibn Anbari JiO Jj

aJlcj J^t '(.Li! Jbojf ^ j»l3, 6 the role of a tertium compara-

tionis {'ilia) was largely formal. That consisted often in the articulation of
a grammatical 'rule', not adduced as justification for the analogy in question,
but rather deduced from its first and second terms and seldom of general
prescriptive value. The restricted usefulness of that kind of reasoning was
enshrined in the Basran dictum cited earlier: l^J* |»jJj ^ iSj*-*
(Jjyij ^JJ- 5 - 7 The analogical foundations of (all) grammar were

not thereby shaken, though it is worth noting that those had, in the face
of pious objections that the word of God was unique, from time to time to
be reasserted: . . . «df j^JI jp*^

ajSjf JJUJI ^ J^.f jJlkj Vj j^I jjCf Xii ^Ul! jSC'T jyi. 8 In its

1 See below, pp. 208-12. 1 Schacht, Origins, no, 117, 125.

3 Goldziher, Zdhiriten, 1 1-12, 41-3, 56, 91-3 : Dawud al-?ahiri, d. 270/884.

4 Cf. Tahanawi, Istildhdt, 1192-5; Tyan, 'Meihodologie', 79-109, esp. 92 ff.

5 For similar tendencies in the application of Talmudic middot, see Loewe, 'The "plain"
meaning', 152-4, and Bacher, Terminologie i, 110-11.

6 Weil, Schulen, 19 n. 3. ? See above, III p. 101.
8 Ibn Anbari, cited Weil, Schulen, 29 n. 1.

PRINCIPLES OF EXEGESIS

169

masoretic form grammatical analogy consisted essentially in the process
of textual restoration/emendation called ultimately taqdir, but in earlier
stages also majaz. 1 The manner in which potentially arbitrary applica-
tion of majdzjtaqdir was restricted by the formulation of grammatical
norms illustrates perfectly the emergence of masoretic analogy and its
conformity with halakhic standards. 2 But that conformity was always
relative, and reflected as much of bewilderment at scriptural grammar
as of piety before the word of God. Both sentiments found expression
in the comment of Ibn Munayyir on Zamakhshari ad Q. 6: 137, that
scriptural data were to be given preference over the grammatical norms. 3
Formal protest against the use of analogy in scriptural exegesis is thus
not unexpected, but an example of such adduced by Suyuti from the
specialized material of Quranic readings and attributed to Dani (d. 444/
1053) may be thought vitiated by the latter's employment of analogy
(qiyas) in his own work. 4

A type of analogical reasoning found in both halakhic and masoretic
exegesis is that based upon the relation between general and particular
statements (khdss:'dmm or kkusus :'umum). Made theoretically complex
by dispute about whether all propositions were primarily/exclusively of
general or of particular significance, application of the principle was in
practice easy. s Apparently general statements in scripture like 'I am the
first of the Muslims' (Q. 6: 163) and 'I am the first of the Believers' (Q. 7 :
143) were interpreted by Ibn Qutayba as particular, on the grounds that
'first' was of temporal value (al-azowalfi zamanihi). b The argument was of
course doctrinal (there had been Muslims and believers from the very
beginning of time), but the same reasoning could be applied to textual
problems, such as the often wayward treatment of number and concord
in scriptural grammar. 7 The corresponding Talmudic precept (kelal:
perat) was of similar application. 8 More often than not, analogies based on
this principle depended upon recurrence of the same word or phrase.
Thus Suyuti, in an argument for the general applicability of particular
Quranic verses, adduced as proof the prophet's juxtaposition of Q. 6: 82

f »lk> p^LjJ pJj and 31: 13 j^kc ^lk! ij^^JI o]. 9 Their common

element, the word zulm (wrong/sin), was designated nazir (analogue):
it is that term which, together with shabah, characterized the masoretic as

1 Wansbrough, 'Periphrastic exegesis', 247, nn. 1-2.

2 See below, pp. 219-24.

3 Kashshdfii, 69-70 (commentary); see below, pp. 223-4.

* Itqdn i, 21 1 ; Dani, Taysir, 21, 22, 34, 128, etc. in the question of phonological assimi-
lation (idghdm), without tertium comparationis, for which however cf. Itqdn i, 214-15, and
GdQ iii, 154.

5 See Goldziher, Zdhiriten, 120-4; Schacht, Origins, 56, 123.

6 Ta'wil, 217. 7 Wansbrough, 'Periphrastic exegesis', 261 nn. 51-2.
8 Bacher, Terminologie i, 79-82, 152-3, ii, 83-5, 161-2. 5 Itqdn i, 86.

170

QURANIC STUDIES

contrasted with the halakhic use of analogy. The utility of nazirjshabah
was most apparent in the solution of textual and grammatical problems by
recourse to the many techniques of restoration and emendation (majdzj
taqdir), but was of course not restricted to such. An example of its applica-
tion to what became a point of doctrine may be seen in the standard inter-
pretation of umm al-kitab in Q. 3 : 7. 1 Descriptions like 'basis' or 'nucleus'
(asl) or 'that which is common to all divine revelations' required, as link to
the 'preserved tablet' of Q. 85 : 21-2, the corroboration allegedly found in
the two other Quranic occurrences of umm al-kitab (13 : 39 and 43 : 4). Now,
that kind of analogy (nazir) could founder on contextual dissimilarity, and
that such did not go unnoticed may be inferred from the several exceptions
to the standard interpretation which identified umm with 'precept' and
'authority'. Those were the tnuhkamat, in reference to which all exegesis
was justified.

3. HALAKHIC EXEGESIS

Among the several topics treated by Muqatil b. Sulayman in his Tafsir
al-khamsimiat aya was the obligation to wage war on God's behalf against
His enemies. 2 That section consists of eighteen Quranic passages (contain-
ing twenty-nine verses) related in the following pattern to six themes:

1. Divinely imposed obligation to fight (qital): Q. 2: 216, zz: 39-40, 9: 29, 49:
9-10

2. Reward for fighting on behalf of God: Q. 61: 4, 10-13, 4: 95-6

3. Observing God's covenant ('ahd) : Q. 9: in, 4: 74, 3 : 200

4. Martyrdom and its reward: Q. 2: 154, 3: 169-70

5. Divine aid against the enemy: Q. 8: 15-16, 65-6, 3 : 155, 9: 25

6. Division of spoils: Q. 8: 41, 3 : 161-3

A degree of thematic overlapping, eliminated from this diagrammatic
exposition, put Q. 3 : 200 after 3 : 169-70, and Q. 9 : 29 and 49: 9-10 at the
end of the section, where the introductory theme was appropriately given
final mention. Muqatil's method can hardly be described as systematic or
thorough: omitted were not only juridical questions traditionally asso-
ciated with the subject of Holy War (jihad), e.g. safe conduct (amdn), but
also a number of Quranic loci pertinent to the themes which he did adduce,
e.g. covenantal obligations. 3 It may none the less be inferred from his
organization of the material that the halakhic theme had priority over the
scriptural evidence marshalled in its support. Selection, and especially
assessment, of the latter were to some extent arbitrary. Introduced by a
tanzil formula, the decree sanctioning war was expressed in terms stressing
contrast with its earlier prohibition, chronologically separated by the Hijra

1 See above, pp. 153. * See above, pp. 163-4; MS BM Or. 6333, 93 v -8 r .

3 See above, I pp. 8-12.

PRINCIPLES OF EXEGESIS 171

same context provided an opportunity to fix tawhid (profession of faith),
salat (ritual prayer), and zakat (voluntary almsgiving: ghayr muwaqqat) as
duties imposed upon Muslims during the Meccan (!) period. The
graduated and contrasting rewards for those who participated in Holy War
(mujdhidun) and those who did not (qd'idun) were modified by a distinc-
tion, not further specified, between non-participants who were excused/
exempted {ma'dhur) and those who were not. 2 The passage setting out
the rewards for martyrdom was inserted into the framework of a three-
fold address from God to the recipients of His favour, culminating in
the familiar story of their desire to return to the world in order to be
killed anew. 3 Verses attesting to divine assistance in battle and, con-
versely, to its withdrawal, e.g. Q. 8: 65-6 and 3: 155, were related an-
ecdotally to the battles of Badr and Uhud, and Q. 9: 25 of course
(textually) to Hunayn. 4 In the passage describing division of spoils the
rule for allocation of the prophet's fifth after his death was not only
enhanced by, but also seen to derive from, an amicable interview between
'A'isha and 'All b. Abi Talib. 5 The style of the whole is unmistakably
haggadic, characterized by the serial repetition of explicative elements and
by a profusion of anecdote. Both devices serve to create an atmosphere of
narratio, so far as such was possible in the thematic arrangement of
material, and the result may be compared with the style of the Sira in
the story of Ja'far b. Abi Talib and the Najashl. 6

Muqatil's use of scriptural shawahid gains significance by juxtaposition
with the treatment of Holy War in the nearly contemporary Muzoatfa' of
Malik b. Anas (d. 179/795). 7 There, the relevant section ('K. Jihad') con-
tains twenty-one chapters related to four themes:

1. Desire {targhib) to wage Holy War: chapters 1, 17-19: Q. 99: 7-8, 3: 200

2. Conduct in Holy War: chapters 2-5

3. Division of spoils: chapters 6—13, 20, 21: Q. 16: 8, 8: 60

4. Martyrdom and its reward: chapters 14-16, 21

The four Quranic passages (containing five verses) appear here not in the
role of organizing principle, but as an almost superfluous embellish-
ment. Q. 99: 7-8 is paraenetic, quite unrelated to the subject, and was
adduced by means of a prophetical hadith beginning: Nothing else was
revealed to me on that matter {scil. instilling a desire to wage Holy War)
save for the general admonition (aya jamia fddhdha) 'Whoever does an

1 See above, pp. 141-2; Tafslr, 93 v . 2 Tafsir, 94 v .

3 Tafsir, 95*; see Wensinck, Handbook, 148 for references in the hadith literature; see
above, p. 125.

* Tafsir, 9 6 r - v . s Tafsir, 9 7 r .

6 See above, pp. 129-30, 133-4; and I pp. 38-43.

' GAS i, 457-64; Muzvatta', 443-71 : 'Kitab al-Jihad'.

172

QURANIC STUDIES

atom's weight of good shall see it, and whoever does an atom's weight of
evil shall see it'. 1 In the same chapter, Q. 3: 200, which recommends
perseverance, is rather more appropriate to conditions of battle, and was
adduced for the same purpose by Muqatil. 2 In chapter 12, on the division
of spoils according to men and mounts, Q. 8 : 60 is but generally pertinent,
while 16:8 was employed to support a halakhic subtlety: whether in the
allocation of plunder donkeys, mules, pack animals, and nags qualified as
mounts. 3 In addition to these four passages, only one of which can be
regarded as serving a juridical purpose, Malik referred twice to 'revelation'
as a source of authority: in chapter 10, Ibn 'Abbas refused to specify
further the anfdl 'mentioned by God in His book' (presumably the hapax
legomenon at Q. 8: 1); and in chapter 14, the prophet supported his own
view in a discussion about the rewards of martyrdom by asserting: And
thus I was told by Gabriel. 4 In neither case was a serious appeal made to
the text of scripture. Moreover, in chapter 10, where the battle of Hunayn
was adduced as precedent, no mention was made of its only Quranic
occurrence (9: 25), employed by Muqatil for another purpose. 5 In chapter
4, two utterances of Malik on the honouring of safe conduct (amdn) did
not include reference to what became the locus classicus (Q. 9 : 6). 6 Now, it
might be argued that neither Q. 9 : 25 nor 9 : 6 is strictly relevant to the
juridical points argued by Malik, but consideration of the entire section on
jihad tends to strengthen an impression that the role of scripture as witness
to correct procedure was indeed minimal. The tendency towards Tslamiza-
tion' discerned in the Muwatta' by both Bergstrasser and Schacht can only
refer to an effort to situate as early as possible the constituents of surma,
not to find corroboration thereto in the text of scripture. 7 The earliest
evidence of the latter is to be found not in the Muwatta', but rather in the
Tafsir al-khamsimi'at aya attributed to Muqatil.

The only extant recension of that work is ascribed to Hudhayl b.
rlabib (d. after 190/805), responsible also for the only version preserved
of MuqatiPs major Tafsir.* In an introduction typically haggadic the essen-
tial components (arkari) of Islam were summarized by means of a parable

related on the authority of Muqatil himself: j***- ^s. jj jJJll* J Is
■iJliJI J| jU LU «uU jlj ojWI ^ Jl jU UJ^ U;

1 Muwaffa', 445. 2 Muwatta', 446. 3 Muwatta', 447.

* Muwatta', 455 and 461, respectively.

5 Muwatta', 454-5 ; cf- Schacht, Origins, 70-1, 286. 6 Muwatta', 448-9.

7 Bergstrasser, 'Anfange', 76-80; Schacht, Origins, 283-7, 3 1I_I 4-

8 See above, p. 144; the isndd of MS BM Or. 6333 (i v ) consists of the last five links in
the chain of transmission for MS H. Hiisnu 17 (i v ).

PRINCIPLES OF EXEGESIS 173

J^J (^jUJI^JI jU LU' -jjU. op £*M o* J-^ J' j^-

f^Jaj ^ ^J* JUaJI J-**^ ^LJI iJI jU LIT JuU op «r**JI jj*-
5i>JI ijl jU. Ij^*. 1 The imagery generated by the concept of a purga-
torial bridge (here jisr and qantara, but elsewhere also sirdt), separating
each soul from its destiny by a series of trials (traditionally seven), was
somewhat unstable. 2 In Muqatil's seven stages: (1) Faith, (2) Prayer, (3)
Almsgiving, (4) Fasting, (5) Pilgrimage, (6) Lesser Pilgrimage ('umra), (7)
Wrongs (mazdlim), it may be that the sixth exhibits contamination with
the preceding one and might well, in the context of the whole work, have
been instead Holy War (jihad). In any case, the subsequent literary history
of this cautionary tale was such that its employment here deserves notice. 3
The treatise covers in fact rather more than the material of the seven
rubrics contained in the introductory parable, though these were given
first places:

Faith (fols. i v -2^), Prayer (fols. 2 v -i2 v ), Almsgiving (fols. i3 r -2i v : including
zakat and sadaqa), Fasting (fols. 2i v -25 r ), Pilgrimage (fols. 25 r ~33 v : including
hajj but not 'umra), Wrongs (fols. 34 r -4i r ), Testaments (fols. 4i r -44 T ), Miscel-
laneous (fols. 45 r ~49 v : including prohibition of usury and of wine), Marriage
(fols. 50 r -S9 v ), Divorce (fols. 6o r -72 r ), Adultery (fols. 72 v ~77 v ), Miscellaneous
(fols. 77 v -93 r : including thefts, debts, contracts/treaties, sacrifice), Holy War
(fols. 93 v -98 r ), Miscellaneous (fols. 98 r -i03 r : including informal prayer and
'contradictions' in scripture, the latter as an appendix). 4

Treatment of each topic conforms with that described above for Holy War:
concepts which are essentially, or even potentially, juridical are presented
as ethical categories, exemplary and hortatory, but rarely prescriptive.
The scriptural loci probantes are tentative and experimental: forming the
principle by which material appears to have been included, but not that
by which its halakhic validity was demonstrated.

Muqatil's ethical categories are rudimentary : halal wa-haram, glossed
by the parable of the purgatorial bridge. The antithesis is itself found in
scripture, formulated negatively (Q. 10, 59, 16: 116), but in the exegetical
tradition positively, e.g. as two of the seven ahrtif (here modes), 5 and in the
earliest literature as gloss to muhkamdt. 6 The five legal categories of classical
jurisprudence were later, and do not appear to represent merely elaboration

• Tafsir, MS BM Or. 6333, i v .

2 Cf. references in Wensinck, Handbook, 40; id. Creed, 232-3.

3 See Asfn Palacios, Escatologia, 180-91, esp. 1S1, 183 (where according to Ibn 'Arabi
the sixth bridge was neither 'umra nor jihad, but toudu: ablutions), and 568-9; Cerulli,
Libro della Scala, 202-5 (paras. 192-3), 299 (paras. 208-9), 530-2.

♦ See above, pp. 164-5.

5 i.e. as two of the seven 'types' of material included in revelation : 'permission* and
'prohibition'; cf. Ibn Qutayba, Ta'wil, 26; Suyuti, Itqdn i, 136; Goldziher, Ricktungen,
37-

6 See above, pp. 149-51.

174

QURANIC STUDIES

of the basic opposition halahharatn, which did not as such achieve herme-
neutical status. 1 As ethical categories halal wa-haram may be compared to
Talmudic miB and "TtOX, or to the related Pauline two? SiSapjs. 2 Their
value for the earliest Muslim exegesis lay in the facility with which
they could be directly and unambiguously applied to the text of scripture,
a procedure which in its most unsophisticated form can be observed in
Muqatil's treatise.

Comparison of the author's method with that of Malik reveals a differ-
ence which eventually became an opposition, namely, Sunna vs. Qur'an
as source of law. The dates of Malik, of Muqatil's ram Hudhayl b.
Habib, and of Shafi'i, with whom the opposed tendencies found polemical
expression, make the end of the second/eighth century a likely chrono-
logical focus for the dispute. It was the merit of Schacht to demonstrate
the crucial role of Shafi'i in that dispute, though I am most reluctant to
accept that the ahl al-kalam 'had a precursor in the author of the dogmatic
treatise ascribed to Hasan Basri', or that the evidence of 'problems which
were based from the beginning on the Koran' proves beyond reasonable
doubt existence of the canonical text of revelation. 3 On the other hand,
Schacht's description of 'Koranic legislation' (sic) as 'the essentially ethical
and only incidentally legal body of maxims contained in the Koran' is in my
opinion especially felicitous, as is his observation that 'Even as regards his
questions which presuppose the rules given in the Koran, we notice that
anything which goes beyond the most perfunctory attention given to the
Koranic norms and the most elementary conclusions drawn from them,
belongs almost invariably to a secondary stage in the development of
doctrine'. 4 In the light of those statements it would not, I think, be un-
justified to interpret references to the 'Koran* throughout his book as of
essentially polemical connotation, employed, as they for the most part are,
in contexts describing Shafi'i's quarrels with his contemporaries and pre-
decessors. 5 In Shafi'i allusions to the Qur'an tend to be perfunctory and
usually in tandem (fil-qur'dn toal-sunna), which suggests a formal hendiadys
alluding to a single source of law, scil. revelation (consisting of both Qur'an
and Sunna). 6 The sunna elevated to the status of revelation was of course
the prophetical Sunna, not the 'living tradition', and one might be par-

1 Cf. Schacht, Origins, 133: post-Shafi'i; id. Introduction, 120-3: designated 'religious
qualifications'.

2 See Gerhardsson, Memory, 303-5 : for Romans 6: 17, and 309, 313; cf. references in
Jastrow, Dictionary, 98, 349, 946.

3 Origins, 224-5, esp. 224 n. 2; see above, pp. 160-3, and I, pp. 44-5.

4 Origins, 224-7, 101 : examples are found on 193-8, 204, 208, 210-13, 215, 218, 250-1,
276-8, 279-80.

5 The references to Qur'an, Origins, 2, seem to me to be ambiguous.

6 Schacht, Origins, 12, 14, 18, 19, e.g. 16: hikma as sunna, 135: 'the two sources'
(aslan) and Shafi'i's 'lip-service to the overruling authority of the Koran, which he did
not recognize in practice' ; and cf. Shafi'i, Risdla, 106-46, tr. Semaan, 'Al-Nasikh', 1 1-29.

PRINCIPLES OF EXEGESIS

175

t doned for asking just what evidence there is for supposing that prior to
' Shafi'i 'revelation' meant exclusively the canonical text of scripture. 1
Opposition to Shafi'i was heterogeneous and widespread, and an
important, but by no means the only, element drew its arguments from the
text of scripture. 2 Aphoristically formulated arguments like 'the book
explains everything' (e.g. Q. 6: 154, 12: in, 16: 89, 17: 12), or 'thus speaks
scripture' and 'any argument not based on scriptural proof is fallacious',
belong to the imagery of polemic and are not likely to have been uttered
from positions unchallenged, or unless pleading a special case. 3 What
seems to be a merely formal recognition of scripture as source of authority:
'what is and/or is not found in Qur'an and Sunna', was characteristic not
only of Shafi'i but also of his predecessors, e.g. Abu Yiisuf, and the same
formula was employed to justify recourse to analogy ! 4 At that stage in the
development of halakhah resort to strictly textual exegesis was rare indeed,
e.g. a varia lectio ad Q. 65 : 6. 5 The use of scriptural passages like 'There is a
fine example (uswa hasana) for you in the Messenger of God' (Q. 33: 21)
may be understood as nothing more than the obverse of 'scripture ex-
plains everything' argumentation, both of which exhibit dispute about
sources, not merely methods. 6

Now, it has more than once been found convenient to distinguish
between the textual relation of law to scripture and their historical relation
to one another. In studies of the Judaic tradition such distinction per-
mitted the assertion that 'In many cases it is quite certain that the Halakhah
antedates the scriptural proof by which it is propped up'. 7 In the Muslim
tradition a parallel distinction was often and intentionally obliterated by
secondary and pseudo-historical conclusions of the kind noticed above in
Muqatil's dating of the divine decrees respectively for Holy War, the
Profession of Faith, Ritual Prayer, and Almsgiving. 8 This version of
the matter, dependent upon a chronology generated by the story of the
prophet's exile (hijra) from Mecca, reflected a working premiss not unlike
that expressed in the Talmudic formula TOO TWKb frsVn." But while the
paideutic function of that and related premisses recommended their
employment in the description of Islamic origins, the haggadic version
proved ultimately something of an embarrassment. The reactions of the

1 Schacht, Origins, 58-81, 149: ShafiTs ignorance of that particular tradition equating
Sunna and Qur'an is hardly relevant, save possibly as evidence that connection of both
with the prophet required still to be articulated; see above, I pp. 51-2, II pp. 56-7.

2 Schacht, Origins, 224, 258-9, but also 40-52: for several significant sources of 'non-
Quranic' opposition.

3 See above, p. 162.

4 Schacht, Origins, 28-30, 101-6, 119, 122; and see above, pp. 166-7.

5 Schacht, Origins, 225, 231-2. 6 Schacht, Origins, 34, 53, 253-4.
' Strack, Introduction, 10; cf. also Gerhardsson, Memory, 83 and the references nn. 1-2.

8 See above, pp. 170-1: Tafsir, MS BM Or. 6333, 93 v .

9 Bacher, Terminologie i, 42, ii, 54-5 ; see above, II pp. 56-7.

176

QURANIC STUDIES

halakhists were summarized by Suyuti in the form of two complementary
principles: material of which the regulative content was effective prior to
its revelation (<u5Cs» ^ <U jy ^li' U) and material revealed prior to its
becoming effective (-0 jy ^ «u5C». ^iti' L). 1 The halakhic relevance of

the latter principle was secondary to its historical function in establishing
a chronology of revelation, by means of which the arbitrary assignment of
Quranic verses respectively to Mecca and to Medina might be lent a degree
of consistency, if not always of plausibility. Thus, the scriptural props for
decrees relating to Almsgiving, Fasting, Holy War, and Ritual Prayer were
regarded as belonging to the earliest stages of Muhammad's prophetical
experience (scil. Meccan), though their regulative content (hukm) was not
enforced until the Muslim community had been founded at Medina.
Revelation {nuzut) of that sort was described as containing a divine promise
(wa'd), and may be compared to evidence of the prophetical vaticinatio in
the Muhammadan evangelium, e.g. Q. 30: 1-4. 2

Rather more important for juridical purposes was the first principle,
according to which ordinances already established and effective were
ratified by revelation. Examples adduced by Suyuti were Ablutions and
Almsgiving, both of which were known and practised before their Quranic

attestation: jTyi (<^5"jJ!) aJ ^Jj U^j<-« dUS J~i» l^i^*^

The operative terms are malum, matluzvj tilawa, and ta'kid, which appear
respectively to signify: promulgated/published (made known), articulated/
articulation in scripture, and corroboration/ratification. Not specified here
are the source and mode of legislation prior to articulation in and ratifi-
cation by scripture. Of such material the greater quantity by far never was
articulated in scripture (ivahy matluw), but was none the less regarded as
revelation (wahy ghayr matluwjwahy marwiy). The distinction between
Quranic and non-Quranic revelation is one to which I have several times
alluded: both were the word of God {kalam allah) and, hence, of identical
authority. 4 The distinguishing element itself was purely formal : tilawa is a
synonym of qur'dn, in the generic sense of recitation. 5 The term might be
used of recitation in prayer, as mode of delivery, and by antonomasia of the
canonical revelation. For the halakhists there was and could be no material
difference between that which was recited as qur'dn and that which was
transmitted as sunna of the prophet. Tilawa was in fact a reference to
'status as scripture', that which with the canonization of revelation could
in fact be found in the document, and may not be interpreted as indication

1 Itqdn, naw' 12: i, 104-6.

3 Itqdn i, 106 citing Ibn Hisar.

5 See GdQ i, 258, iii, 144 n. 5.

2 See above, pp. 144-5, and II pp. 69-70.

4 Suyuti, Itqdn i, 137-8, iv, 174.

PRINCIPLES OF EXEGESIS

177

of contrasting modes of transmission (oral and written) analogous to
Rabbinic terminology. 1 The significant parallel between the Judaic and
Muslim traditions was insistence upon a single source of legislation, which
was divine.

Verses of regulative content, but not including the many non-prescrip-
tive passages which might be adduced as loci probantes by the halakhists,
make up approximately one-sixth of the Qur'an. More than half of these
are found in suras 2-9, though material of potentially legal application is
scattered throughout the book in no discernible pattern of distribution.
Explicit reference to the commandments of God {hudud alldh) is rare and
unsystematic, e.g. Q. 2: 187 (fasting), 2: 229-30 (divorce), 4: 13-14
(verses 1-12 concern testamentary matters), 9: 97 (on the treachery of
bedouin in contractual obligations), 9: 112 (paraenesis, but related to the
foregoing), 58: 4 (verses 1-4: divorce), 65 : 1 (divorce). The technical term
hadd in penal law is thus only symbolically related to the scriptural hudGd. 2
More important than hudud, and not restrictecLto penal law, was the
hermeneutical concept muhkam, related from the time of Muqatil to the
general prescriptions in Q. 6: 151-3 and by Maturidi to both Q. 6: 151-3
and 17: 22-39.' Designation of these two passages as the Quranic 'Deca-
logue' 4 is thus not quite so fanciful as Obermann appears to have believed,
though a verse-by-verse correspondence is not really justified. 5 But if the
principle contained in muhkam offered a theoretical point of departure for
both halakhic and masoretic exegesis, the finding of specific and useful
juridical material (hukmjahkdm) in the text of scripture was in practice
frustrated by the absence of an unambiguous and uncontradictory historical
framework. Solutions to the problems resulting from that condition were
sought, and for the most part found, by imposing upon the document of
revelation a chronological stencil. Historical order could thus be intro-
duced into what was essentially literary chaos.

To that end the primary device employed was description of the circum-
stances of revelation (asbdb al-nuzul, but also mawatin, awqdt, waqiat,
akhbar). I have touched upon the incidence of its haggadic application, in
which concern for the narratio was paramount. 6 Elaboration and refine-
ment of the technique were the work of the halakhists. An instructive

1 See Goldziher, Studien ii, 194—202; in Rabbinic terminology stress was anyway on
mode of delivery rather than of transmission, see Strack, Introduction, 12-20.

2 Cf. Schacht, Origins, 126, 191, 208-10.

3 Muqatil, Tafs.tr, MS H. Hiisnu 17, 35'; see above, p. 149; Maturidi, MS Medine
179, n6 v -77 v .

+ e.g. Hirschfeld, Researches, 81-2; Speyer, Erzdhlungen, 305-10; Goitein, 'Birth-hour',
in Studies, 132; Katsch, Judaism, 152.

5 Obermann, 'Agada', 38 n. 2; to contend that the only Quranic occurrence of cove-
nant (mithdq) with allusion to B. Israel is Q. 2: 83 is arbitrary and irrelevant: the Mosaic
context of both Q. 6: 151-3 and 17: 22-39 is quite clear.

6 See above, pp. 141-2, and I pp. 38, 41.

178

QURANIC STUDIES

summary of that process may be read in Suyuti. 1 A considerable portion of
the discussion was devoted to the precision of temporal and spatial occasions
of revelation, from which emerges an unmistakable impression of arbitrary,
if not irresponsible, assignment. 2 For most of the loci probantes there
adduced, non-regulative and hence halakhically neutral, the epithets
'Meccan' and 'Medinan' were not even mutually exclusive. For regulative
material the contrast specific : general (khdss:'dmm) proved of some value
in distinguishing just which of many possible verses represented the first
and particular enactment of a decree {awail makhsiisa), e.g. for Holy War
(qitdl), dietary laws, prohibition of wine, etc. 3 And further differentiation
was available. Quranic revelation was alleged to be of two kinds: spon-
taneous (ibtidd'an), or in response to an event or a query ('aqiba waqi'a aw
su'dl). Application and reference of the latter were not, however, limited to
the particular event or query which had inspired them: the operative
distinction was found to lie between particularity of cause (khusus al-sabab)
and generality of expression ('umum al-lafz).* But the arbitrary character
even of this ruling becomes apparent in a discussion of the elative al-atqd in
Q. 92: 17 'He who is (the) most pious shall be spared'. Desire to restrict
that reference to Abu Bakr provoked some very dogmatic observations on
the grammatical function of the definite article. 5 The agreed general
chronology of revelation (twenty to twenty-five years divided approxi-
mately between Mecca and Medina) generated yet a further distinction
between 'cause of revelation (sabab) and 'report about' revelation (khabar) :
an example was Sura 105 and the story of God's protection of the Meccan
sanctuary. 6 Even that technique, which could be applied to all Quranic
data on God's earlier interventions in history, was susceptible of modifi-
cation: the reported miracle must be seen to have a cause, which was God's
bestowal of His word upon His prophet, whether or not the latter was
identified. The khabar was thus also a sabab. 7 Moreover, a single verse
might have had several causes/occasions of revelation, e.g. both Q. 9: 113
and 16 : 126 could each be traced to three separate events, well spaced in the
career of Muhammad and hence both Meccan and Medinan: ^
JjjJ) ^OjC^ ii.pUVI Recognition of that possibility was naturally

exploited to explain the phenomenon of repetition (takrar) in the docu-
ment of revelation. 8

Now, it ought to be clear from this summary of methods pertinent to

1 Itqdn, ataxia' 1-15 : i, 22-1 15, drawing extensively upon the classical work of Wahidi
(d. 468/1076), GAL i, 411-12, Suppl. I, 730-1, Kitdb Asbdb al-nuzul.
* Itqdn, amvd' 1-8: i, 22-81 ; see above, pp. 126-7.

3 Itqdn i, 74-6. * Itqdn i, 82, 85. 5 Itqdn i, 87.

6 Itqdn i, 90; see above, I pp. 42-3.

7 See 'Abd al-Jabbar, Tanzih al-Qur'dn, 480; and above, II pp. 73-4.

8 Itqdn i, 95-6, 102-3.

PRINCIPLES OF EXEGESIS

179

^ Quranic chronology that the historical value of such material is restricted.
A single criterion, articulated by Wahidi, found general acceptance: valid
reports about the occasions of revelation must be based upon eye-witness
accounts. 1 With that statement the entire subject was subsumed under the
general rubric Tradition, and its expression can be assessed only by
reference to the standards generally obtaining for evaluation of hadith
literature. Production of that literature rested upon two convictions: that
the reliability (thiqa) of a witness could be known, and that for such
continuity of transmission (isnad) could be established. The manner in
which that was accomplished in the field of legal traditions was described
by Schacht. 2 That so-called 'historical' traditions came into existence in
precisely the same way as legal ones is clear from examples like those
pertaining to the marriage of the prophet to Maymiina. 3 It seems at least
doubtful whether for exegetical (tafsir) traditions a different origin can be
claimed. That these exhibited a reaction to the undisciplined employment
of subjective criteria in scriptural interpretation was proposed by Gold-
ziher. 4 From a purely formal point of view there is something to be said for
that proposal, but from the point of view of substance, it may be observed
that exegesis provided with formal isnads can rarely be distinguished from
that without. The supplying of isnads, whether traced to the prophet, to his
companions, or to their successors, may be understood as an exclusively
formal innovation and cannot be dated much before 200/815. That
Shafi'i's stringent standards with regard to prophetical hadiths were not
applied in the fields of history and exegesis is an impression derived from
a wholly artificial classification of their contents. The substance of history,
of exegesis, and of law was identical: its degree of attestation depended
upon the particular use being made of it. And the quality of isnad (marfu,
muttasil, mursal, maqtu', etc.), too, varied for the same material according
to its employment. 5 The frequently adduced view that the text of revelation
was easily understood by those who had witnessed its first utterance, as
well as by their immediate successors, but by later generations could not
be, is in my opinion not merely ingenuous, but belied by the many stories
of early efforts towards the interpretation of scripture associated with the
figures of 'Umar b. Khattab and 'Abdallah b. 'Abbas. Whatever the reasons
for production of those stories, it seems hardly possible that at the begin-
ning of the third/ninth century the Muslim community had to be reminded
of what it had once known. Tafsir traditions, like traditions in every other
field, reflect a single impulse: to demonstrate the Hijazi origins of Islam.
The earliest extant work on the circumstances of the Quranic revelation

1 Apud Suyutf, Itqdn i, 89. 2 Origins, 36-9, 163-75.

3 Origins, 138-40, 153.

4 Richtungen, 61-5 : tafsir manqicl (bil-'ilm) as against tafsir bil-ra'y.

5 Pace Horst, 'Zur Oberlieferung', 305-7 ; and cf. Birkeland, Muslim Opposition, 28-42.

QURANIC STUDIES

is ascribed to Ibn Shihab Zuhri (d. 124/742) and entitled Tanzil al-
Qur"an. 1 The ascription is arbitrary, but need not mislead: the recension is
that of the Sufi exegete Sulaml (d. 412/1021) 2 and in complete accord with
the later accepted tradition on the chronological order of suras, shorn of the
subtleties appropriate to scholastic discussion of the subject, and even of
recognition that a single sura might contain material of both Meccan and
Medinan origin. On that particular point the author was revealingly

explicit: a5Cj c^p a5v*j ojj*» cJji lit j^j- 3 The number of suras
assigned to Mecca is 85 and to Medina 2c;, 4 and the internal order corre-
sponds to that of SuyQti's third list. 5 For the history of Quranic exegesis
this bald statement of fact circulated in the name of Zuhri is quite with-
out value. The isnad is anyway defective and the last authority but Zuhri,
one Walid b. Muhammad Muqari, was considered matruk al-hadith. 6

More important for the study of both tafsir traditions and asbdb al-nuzul
is the 'Kitab Tafsir', included as-ehapter 54 in the Sahih of Muslim (d.
261/875). 7 That very brief treatise consists entirely of witness to the
occasions of revelation for sixteen Quranic verses, traced to the authority of
'A'isha, 'Umar, Abu Hurayra, and Ibn 'Abbas, with general mention of
suras 8, 9, 59, and of the prohibition of wine. Reference to abrogation
(naskh), the only reason for halakhic interest in the chronological order of

revelation, is explicit: Ibn 'Abbas declared that Q. 4: 93 LuJ^ J3j ,v«j
(( ej\y*i IjujOU was the last to have been revealed and had thus not
been abrogated: l^sc^J L> J)jf U jL\ tz^yl Jul. 8 In a discussion
of whether Q. 5 : 3 (dietary laws) had been revealed about the Jews, 'Umar
asserted his authority on the grounds that he knew all of the asbab al-
nuzul. 9 Ibn Mas'ud dated the revelation of Q. 57: 16 by reference to his
own conversion four years earlier. 10 Muslim's material contains the
premisses but not the arguments of halakhic exegesis: the chapter is
fragmentary and badly organized, and may owe its very existence to the
author's recognition that in a collection of traditions a few on the subject
of scriptural exegesis would not be out of place. From that admittedly
conjectural reading of the evidence one might conclude that for tradition-
ists, even after the canonical text of revelation had been established, the
Qur'an was merely one topic among many requiring the formal embel-

1 GAS i, 283. * GAS i, 671-4. 3 Zuhri, Tanzil al-Qur'dn, 32.

4 Duplication of Sura 7 and absence of Sura 33, copyist's errors, are remarked by the
editor, 30 n. 1.

5 Itqdn i, 26-7 ; reproduced GdQ i, 59-60. 6 Cf. Goldziher, Studien ii, 144.

7 GAS i, 136-43; Safilli viii, 237-46.

8 Muslim, Sahth viii, 241 ; adduced by Suyuti, Itqdn i, 80, among a number of equally
well-attested candidates for that honour.

• Muslim, 238.
10 Muslim, 243.

PRINCIPLES OF EXEGESIS

181

Iishment of hadith. The hypothesis might be thought corroborated by
the treatment of tafsir in other collections of traditions.

In the Sahih of Bukhari (d. 256/870) the 'Kitab Tafsir' (chapter 65)
occupies a prominent position. 1 In the corpus of 475 traditions every
Quranic sura received some attention from the author, if only in the form of
a simple lexical identification on the authority of Mujahid or Ibn 'Abbas, of
one of the latter's disciples, or occasionally without citation of any author-
ity at all. 2 Bukhari's lexicology could be described either as decidedly
primitive or as presupposing a long tradition in the course of which stan-
dard solutions to major problems had crystallized and no longer required
authentication. Whichever of the alternatives is more likely, the presence of
such material in a collection assembled to demonstrate the importance of
traditional authority does not inspire confidence. 3 A popular etymology for
the name Gabriel, adduced from 'Ikrima ad Q. 2: 97, deserves notice:

<U!l JjI Jl.£. dL^j^o.. 4 This was followed by a tradition from Anas

on the 'rabbinical test of prophethood', here put to Muhammad by 'Abdallah
b. Salam. 5 0.2:31 became a peg for asserting, by means of a Purgatory
motif, the rank in heaven of Muhammad above all other prophets. 6 Ad
Q. 2: 136 Abu Hurayra was cited for the prophet's prescription on the
proper conduct of Muslims towards Jews. 7 Ad Q. 2: 183 a hadith from
'Abdallah b. 'Umar announced the substitution of Ramadan for the earlier
pagan Arab fast (sic) of 'Ashvira'. 8 Ad Q. 17 : 85 definition of the Spirit
(ruh) was related to the Jews. 9 The entire passage on Surat al-Kahf was
devoted to the story of Musa and Khidr (Q. 18: 60-82), with special con-
cern for the identity of Musa. 10 Bukhari's exegetical method was, in brief,
predominantly haggadic, and the absence of anecdotal material for such
popular passages as Q. 30: 1-4, 85 : 21-2, and Sura 105, must be regarded as
fortuitous. The essential difference between Bukhari and the haggadists is
the insertion of appropriate isndds, many of which, however, were carried
no further than to a successor (e.g. Mujahid). The occasional intrusion of
an element specifically halakhic or masoretic may be noted : e.g. whether
the 'compensation clause' with regard to fasting in Q. 2: 184. (fidya fa' dm
miskin) had or had not been abrogated; whether the waiting period ('idda)
for divorced and/or widowed women was regulated by Q. 2 : 234 or by
2: 232; the relative merits of the variants an yattazowaf and alia, yattawwaf
in Q. 2: 158; explanation of lakinna in Q. 18: 38 as lakin and, produced
by a combination of elision (hadhf) and assimilation (idgham). 1 1 Observations

1 GAS i, 115-34; ed. iii, 193-390.
3 See below, pp. 216-19.
5 See above, pp. 122-6.
7 Bukhari, 197.

' Bukhari, 275; see above, pp. 125, 128.
" Bukhari, 202, 207, 200, 276, respectively.

2 Cf . Birkeland, The Lord guideth, 40.

4 Bukhari, Sahih iii, 196.
6 Bukhari, 194-5; see above, pp. 172-3.

8 Bukhari, 201-2 (variant: Quraysh).
10 Bukhari, 277-82; see above, pp. 128.

QURANIC STUDIES

on the occasions of revelation are circumstantial only, and never explicit
as in Muslim's Sahih.

What must be regarded as the raison d'etre of tafsir traditions in the
major collections was most clearly formulated in the 'Kitab Tafsir' of
Tirmidhi (d. 279/892), as chapter 44 of his Sahih. 1 While the material
itself is haggadic and for the most part identical to that adduced by
Bukhari, the chains of transmission are not only more complete, but each
provided with comment on its degree of acceptability (e.g. sahih, hasan,
gharib). Explicit reference to the purpose of that exercise was set out in an
introductory paragraph, in which is stressed the danger of subjective
exegesis {tafsir bil-rdy) and, conversely, the necessity of authoritative
tradition ('ilm). z Coverage of the text of scripture is not complete (unlike
Bukhari), though adequate up to Sura 75 (a total of twenty-one suras was
not provided with comment). For verses containing a crux interpretum,
e.g. Q. 3: 7, alternative isnads were adduced. 3 Verses traditionally em-
ployed as pegs for anecdote, like Q. 17: 1, 18: 60-82, 30: 1-4, were treated
as in Bukhari, perhaps even more generously. 4 References to the asbab al-
nuzul are, as in the latter, merely implicit and circumstantial. Allusion to the
instrument of abrogation is explicit for the qibla controversy, e.g. Q. 2: 115
and 134. 5 Of textual exegesis there is virtually none, save for the variants
an and alia ad Q. 2: 158. 5 Lexical explanation resembles that of Bukhari,
but is sporadic and not, as in the latter, adduced in concentration at the
beginnings of suras. Chapters on thawab al-Qur'an and qiraat in Tirmidhi
(nos. 42, 43), like the chapter on fada'il al-Qur'an in Bukhari (no. 66),
attest to a view of scriptural studies somewhat more sophisticated than that
displayed by their contemporary Muslim, but at the same time exhibit
rather more of cliche and stereotype formulation.

The rhetorical analysis of hadith literature, as contrasted with its
exploitation for legal, historical, and sociological purposes, has attracted
very little scholarly attention. Two studies in particular deserve mention :
Vajda, 'Jviifs e t Musulmans selon le hadit' (1937); and Stetter, Topoi und
Schemata im Hadit (1965). The most significant feature of that literature,
signalled by both "Vajda and Stetter, is its schematic formulation. Employ-
ment of much circumstantial and 'naturalistic' detail, judged also by
Schacht as indicative of fictive situations, 7 tends to fall into recognizable
and even predictable patterns. For example, emphasis upon the pastoral
simplicity of Jahiliyya and early Islam, accompanied by unsubtle humour
at the expense of bedouin manners, attention to every aspect of the
prophet's personality, and recurrent use of expressions implying intimate

1 GAS i, 154-9; ed. xi, 67-xii, 264. 2 Tirmidhi, Sahih xi, 67-8.

3 Tirmidhi, xi, 114-20; see above, pp. 149-53.

4 Tirmidhi, xi, 290-3, xii, 2-13, 66-72, respectively.

s Tirmidhi, xi, 79-88, esp. 80. 6 Tirmidhi, xi, 90.

7 Schacht, Origins , 156.

PRINCIPLES OF EXEGESIS

183

recollection (e.g. Jzi] and J I JUcj UJ) are of such regularity as to
suggest a common pool of narrative ingredients. 1 That circumstantial de-
scription might also contain elements of halakhic value, in particular refer-
ences (explicit and implicit) to time and place, must be acknowledged. But
the very ubiquity of both motif and formula is significant. The hadith litera-
ture reflects both form and substance not only of juridical concern with the
actions and utterances of the prophet of Islam and with the contents of the
Quranic revelation, but also of its haggadic (narrative and historical) ex-
pression in stra, maghazi, and ayyam. The presence of isndds as halakhic
embellishment is, from the point of view of literary criticism, a superfluity.
The substance of Bukharl, Muslim, and TirmidhI is that of Muqatil, Ibn
Ishaq, Sufyan, and Kalbi. It is also that of the entire exegetical tradition,
excluding the masoretic literature, up to and including Suyuti.

A single illustration in place of many: a flagrant tendency discerned by
Vajda in the hadith literature was the transposition of anti-Jewish elements
of Islamic prescription into the category of superseded Jahili custom. One
example was designation of 'Ashiira' not as a Jewish, but as an ancient
Arabian practice. 2 Others were abolition of the custom of public lamenta-
tion at funeral processions and of abstention from sexual intercourse
during menstruation, both identified with pagan Arab practice. 3 The
incidence in haggadic exegesis of that kind of transposition, by which the
roles of Arabian Jewry and Quraysh were exchanged or combined or
otherwise blended, has been described. 4 The ultimate value of the tech-
nique was doctrinal; its origin, however, was polemic, not quite effaced
from the memory of the Muslim community even in the third/ninth
century.

The several ways in which the halakhists employed tafsir traditions may
be seen in three kinds of exegetical literature: ahkam (prescription),
ikhtildf (dispute), and naskh (abrogation). While the scope of each extended
beyond exclusively midrashic exploitation of the text of scripture, it is with
that procedure in particular that I am concerned. The extrapolation of law
from revelation was, in the Muslim community as in others organized on
similar theocratic principles, a tortuous and interminable process. s Ex-
ceptions to the accord and harmony symbolized in the notion of consensus
secured recognition in the complementary notion of permitted areas of dis-
pute. For very few problems was there ever a final solution or even a set of
agreed scriptural references. Prescriptions relating to Holy War are a case
in point. For Jassas the obligation to fight in the way of God was derived

1 Stetter, Topoi und Schemata, 4-34.

2 Vajda, 'Juifs et Musulmans', 122-3; Bukhari, Sahih iii, 201-2; see above, p. 181.

3 Vajda, 'Juifs et Musulmans', 78 n. 1 and 69-75, respectively.

4 See above, pp. 122-7, an< ^ H PP- 62-3, 70-3.

5 e.g. Rabin, Qumran, 82-94, 95-111.

1 84

QURANIC STUDIES

not from Q. 2: 216 (as for Muqatil) but from Q. 2: 190 J.^ J IjJjUj

at m

^XjlJI i_^>l> V -Oil 61 IjJCjJ |*SCijij'l2j ^jJI dJl in which was
stressed the exclusively defensive character of combat imposed upon
Muslims. 1 Precluded by its terms were thus non-combatants, such as
women, children, and hermits/monks (ruhbdnjashab al-sawami'). A different
qualification was that made between infidels (mushrikun) and scriptuaries
(ahl al-kitdb) as liable, by divine decree, to attack by believers. Of the
four verses (Q. 2: 191 and 4: 91, 4: 89 and 9: 5) adduced to support the
progression from defensive to offensive warfare and from selected targets
to a general declaration of hostility to non-Muslims, 0.9:5 became the
scriptural prop of a formulation designed to cover any and all situations
which might arise between the Muslim community and its enemies,
and included lingering compunctions about clauses attaching to the
sacred months (ashhur hurum) and the sanctuary at Mecca (masjid harairi).
Called in the exegetical tradition the sword-verse (ay at al-sayf), Q. 9 : 5

jl |,gl. ; .,r Ijlicj oj5^31 0jL*sJI lj-»lslj IjjU' jj^ JS" ij-^L?

(*s*-J jjic. achieved a quite extraordinary status in the elaboration of

Islamic jurisprudence, as the alleged abrogant of 1 24 Quranic verses. 2 These
included all passages in scripture which could be interpreted as recom-
mending leniency (safhwa-'afw; cf. Q. 2: 109, 5: 13) towards unbelievers.
The range and variety of such were equally extraordinary, at least as set
out in what became the classical work on Quranic abrogation, the Kitab
al-ndsikh wal-mansUkh of Hibatallah (d. 410/1019). 3 The ubiquity of the
abrogant dyat al-sayf in that treatise suggests that it was not the law of
Holy War at all, but the presence of legislative repeal in the text of scrip-
ture which was being argued. 4 But Hibatallah represented the final stage of
halakhic exegesis, in which general conclusions could be advanced without
authorities and shorn of scholastic justification. A century earlier Nahhas
(d. 338/950) 5 had observed that dyat al-sayf, at least with regard to the
treatment of prisoners of war, was itself abrogated by Q. 47 : 4, the view of
Hasan Basri and others. 6 Some, on the other hand, held that the opposite
was true: the leniency of Q. 47: 4 had been abrogated by Q. 9: 5. Nahhas
himself argued that neither had been repealed, that both verses were
muhkamdt (sic, cf. Q. 22 : 52) since they were not mutually exclusive, and
that decision on the treatment of prisoners lay with the imam. That view
was supported by several traditions on the action of the prophet during the

1 Ahkdm al-Qur'dn i, 256-63 : 'Bab Fard al-Jihad'; see above, p. 170.

2 e.g. Ibn 'Arabi, apud Suyuti, Itqdn Hi, 69. 3 GAS i, 47-8.
4 Hibatallah, Kitab al-ndsikh wal-mansukh, 29, 37, 38, 51, and passim, esp. 53-88 for

suras 11—54, most of which could boast only one verse abrogated; see below, pp. 196-7.
s GAS i, 49. 6 Nahhas, Kitdb al-ndsikh wal-mansukh, 165-6.

PRINCIPLES OF EXEGESIS

185

^ conquest of Mecca. Much later Ibn 'Arabi, in whose discussion the general
status of dyat al-sayf as abrogant attributed to him by Suyuti is in fact not
found, employed the instrument of khdssf 'atnm to demonstrate that
although the specific reference of Q. 9: 5 was to pagan Arab idolaters
('abid lil-wathan), the verse was generally valid for all who rejected God's
message, whether during the sacred months or within the Meccan sanc-
tuary. 1

The employment of tafsir traditions in the ahkam literature was not
limited to precision of juridical niceties. For Q. 3: 200 l^uT ^jJI L^jU
OjxbJi> p5XJ 1^21 1 j l_^Jajljj IjjiU-J ljjr*»'> adduced by both Muqatil
and Malik in the spirit of general paraenesis, Jassas adduced two utterances
of the prophet equating ribdt for the sake of God with the virtue deriving
from fasting and prayer. 2 But the primary purpose of such traditions
was halakhic: to render explicit that which was seldom more than
implicit in the text of scripture, by establishing a specific historical con-
text for the revelation in question. In the much-disputed problem of

reference in Q. 5: 33 J Oy^-J <0>"jj <*-Ul dyj^i <ji'-)J\ e.\y>. Uit

Jassas provided two haditks : one from Ibn 'Abbas applying the verse to
polytheists (mushrikiin), and another from Ibn 'Umar identifying 'those
hostile to God and His prophet' as the clan of B. 'Urayna ('Uraniyyun). 3
Summing up the evidence (there are several additional hadiths from Ibn
'Abbas, conflicting and with different isndds) as offering a choice between
polytheists and apostates, Jassas found himself constrained to reject both
on the grounds that whatever the occasion of the revelation, its semantic
content was clear: reference was to all transgressors of God's law. The
principle thus enunciated was that juridical application could not be based
upon an 'occasion' (presumably a historical accident) but only upon the

general validity of the scriptural expression: Ui]j Uj^* i_^JJ ^ <jV

JaiUII Ujcx p5G*J|. 4 The manner in which this kind of argument cut

across earlier exegetical method based on the chronological arrangement of
scripture emerges from a comparison of Jassas with Abu TJbayd, who re-
ported that the incident involving B. 'Urayna had taken place in the
early years of Islam (ft awwal al-isldm) before the revelation of these pre-
scriptions (qabla an tunazzal al-hudad), and that according to Ibn 'Abbas ( !)

1 Ibn 'Arab!, Ahkam al-Qw'dn i, 369-70.

2 Ahkam al-Qur'an ii, 45: the metaphorical extension of that concept, from ribdt
al-khayl to ribdt al-nafs, generated some interesting pseudo-history in medieval North
Africa, see Norris, 'Origins of the Almoravid movement', 255-68, esp. 263-5.

3 Ahkam al-Qur'an ii, 406-8. * Jassas, loc. cit.; see above, p. 178.

186 QURANIC STUDIES

the contents of Q. 5 : 33 left the precise punishment of transgressors to the
discretion of the imam. 1 Now, the chronological relationship between that
particular event and the revelation repealing the prophetical Sunna based
upon it is anything but obvious. Undated in the Sira and fixed at Shawwal
6/February-March 628 by WaqidI, the affair of B. 'Urayna appears to have
been one of several movable feasts in the early Islamic calendar. 2 Its assign-
ment to the earliest period of prophetical activity by Abu 'Ubayd and to a
much later date by WaqidI was evidence of concern not with its historical
truth, but with its juridical value.

The implications of Q. 5 : 33 were never unanimously clarified, though
that could hardly have been for want of effort. As an appendix to the
sections on jihad and jizya in his Ikhtildf al-fuqaha, Tabari adduced in his
treatment of the subject a degree of consensus on identification of the
muharib as sinner (fasiq) rather than as infidel (kdfir), but added that the
protected non-Muslim (dhimmi) was also liable to the ahkam al-muharibtn,
since transgression was tantamount to violation of a treaty. 3 It was further
stipulated that the affair of B. 'Urayna had taken place before revelation of
Q. 5 : 33, and that that was juridically relevant. The sophisticated logic of
Jassas's argument had presumably not yet been formulated. 4 The ahkam
al-muharibin are the subject also of a short treatise tentatively, but in-
correctly, ascribed to 'Ata Khurasan! (d. 136/757).* In fact the work,
contained in seven folios inserted at the beginning of MS Ahmet III, 310, 6
represents a post-Tabarl stage of ikhtildf. Considerable attention was
devoted to the circumstances in which Q. 5 : 33-4 had been revealed, and a
series of traditions adduced, claiming 'some group of ahl al-kitdb who had
broken their covenant with the prophet' (sic), of B. Qurayza when they had
planned to assassinate the prophet, or (unspecified) infidels, or finally, B.
'Urayna. 7 Whether the scriptural passage could be held to have abrogated
the prophetical sunna explicit in the B. 'Urayna tradition depended upon
the identification of muharib (enemy) with murtadd (apostate), an equation
for which corroboration might be found in the reference to tawba (con-
trition) in Q. 5: 34. For the author of that treatise consensus doctorum
seemed to support the identification, and the remainder of the text is

1 See above, p. 150: Kitdb al-ndsikh wal-mansiikh, MS Ahmet III, 143, 94 v -6 r .

2 Cf. Jones, 'Chronology', 253, 279.

3 Tabari, Kitdb Ikhtildf al-fuqahd', 242-9 : an example of analogy by ta'lil, see above,
pp. 167-8.

4 That the proper concern of the lawgiver was with general principles rather than with
specific circumstances (inferable, after all, by analogy and other hermeutic devices), had
been articulated in both Hellenistic and Rabbinic legal literature, see Daube, 'Rabbinic
methods', 247-51.

5 GAS i, 33; MS Ahmet III, 310, 1-7.

6 One of many copies of Qadi 'Iyad, Kitdb al-Shifd': no. 2733 in Karatay, Topkapi
Sarayt Miizesi Kiitiiphanesi Arapfa Yazmalar Katalogu, ii, under which entry this in-
sertion with separate pagination is not mentioned.

7 MS Ahmet III, 310, 2 r " T .

PRINCIPLES OF EXEGESIS 187

^ devoted to defining transgression of the divine law (kudud alldh) as con-
sisting primarily in murder {qatl), theft (saraqa), and banditry (qaf al-
iariq).

Two technical terms employed in Ahmet III, 310, and elsewhere in the
ahkdmjikhtildf literature(s), deserve notice. In discussion of the punish-
ment which could be imposed upon the muhdrib, a problem was created by
the options set out in Q. 5 : 33 : he was to be killed, or crucified, or have his
hands and feet severed, or be banished. The choice might be left to the
discretion of the imam or based upon a precise definition of the culprit's
offence. Tabarl was stated to have preferred (murajjih) the latter, and that
view was preferable (rdjih) since it conformed to the text of scripture
(huwa nass al-dyd). 1 On the manner of crucifixion where such was deemed
appropriate, the view of Shafi'i was that the offender should be crucified
alive and then killed on the cross, that being explicit in scripture {huwa
'l-azhar min al-dyd) which recommended an exemplary punishment. 2
Now, in neither instance does connection of rdjih or azhar with the text of
the verse in question indicate an obvious interpretation derived from the
'plain meaning' of scripture. 3 From the syntax of Q. 5 : 33 it was not quite
possible to insist that the series of proposed punishments exhibited
gradation according to the nature of the offence, or that the notion of
'exemplary' punishment (matkula) was dominant. Limitation or exclusion
of the imam's discretion and interpretation oijaza' (recompense) as deter-
rent reflect substantial and sophisticated additions to the wording of
scripture, and may be traced to the story of the prophet's action in the
affair of B. 'Urayna. Employment of the term azhar (or zdhir) was emotive,
of rdjih (or arjahjmarjuh) subjective or at best conjectural. 4 Halakhic use of
tarjih for 'preference of one of several options' was a process justifiable only
by resort to the abiding distinction between muhkam and mutashdbih,
itself postulated as an integral characteristic of scripture. 5 In more general
exegetic usage tarjih was required to conform to the normative standards of
analogy based upon juxtaposition of identical and similar passages. 6 In
Ahmet III, 310 the terms zdhir and rdjih are synonymous and inter-
changeable and, save for the very special usage of the Zahiri madhhab,

such remained the practice in Muslim exegetical literature: J yblklls

I J loJo*^l ilJbd!. 7 Worthy of mention is Maimonides' use of

zdhir, with appeal to the text of scripture (nass), in an exegetical context

1 MS Ahmet III, 310, 3 r , 4 T . 2 MS Ahmet III, 310, 3*.

3 See above, pp. 150-1, 152-3.

4 A point insisted upon by Fakhr al-dln Razi, apud SuyutJ, Itqdn iii, 12.

5 Itqdn iii, 31-2, also citing Razi; see above, pp. 15 1-2.

6 Particularly for the masorah, see Suyuti, Itqdn i, 31, 58, 93-4, 229, and ii, 264, where
a distinction between marjuh and zdhir may be thought scarcely discernible.

' Goldziher, Zdhiriten, 24 n. 4 citing Juwaynl.

QURANIC STUDIES

requiring even more imagination than the allocation of penalties in
Q. 5 : 33, namely, the messianic symbolism of Daniel 7-8 : ~>nbn TIDX rnm
pbx Tli» p. 1 The concept of 'plain meaning' was indeed a generous one.

Whatever the linguistic and logical assertions made about the ipsissima
verba of scripture, halakhic exegesis turned upon the assumption of a
chronological, and hence causal, relation between Qur'an and (prophetical)
Sunna. The question of priority, though hedged with qualification, was
generally answered in favour of the latter. Jassas, who had elsewhere dis-
missed the occasion of revelation as irrelevant to interpretation of the
Quranic text (in the case of B. 'Urayna and Q. 5: 33), admitted that in
juridical matters the specific meaning (takhsis) of the Qur'an could be
determined by reference to a prophetical tradition, provided such was
widely {min tariq al-tawatur) not sparsely attested (bi-akhbar al-ahad). 2 The
particular context of that declaration was the abrogation (naskk) of Q. 7 : 3

^ p5CJI <}-J\ U by 59: 7 ^l&i Uj cjjJ*s J^JI j^UT Uj
I li -up. Resort to Sunna was thus justified by Qur'an, and the con-
clusion drawn that for halakhah the prophet was of the same authority
as the text of scripture: jTpUl |*5C>JI v^i J The locution

'of a status with (=f like) Qur'an' could be misleading: it is quite clear
from the context that it was scripture being subjected to interpretation, an
operation for which the hermeneutical instrument was tradition, in the
form of words and deeds attributed to the prophet. These were preserved
and transmitted with the care prescribed for canonical revelation. 3
References to the contrast litteratim transmission (riwdyajnaql bil-lafz) :
paraphrastic transmission {riwdyajnaql bil-ma'na) may not be understood
as allusion to mutually exclusive modes, but, rather, to a polemically
formulated concern for the authority of the Quranic text. 4 Paraphrastic
transmission was anyway, in the view of the halakhists, limited to utter-
ances from the companions of the prophet {al-sahaba), an attitude which
may indeed have had its origins with, but was not restricted to, the Zahiri
madkhab.* An indication of stringency in the transmission of Sunna may
be seen in Abu 'Ubayd's discussion of the prophet's action in the affair
of B. 'Urayna: the phrase 'he put out their eyes with spikes (iron)' was
preserved in two versions exhibiting the purely phonological contrast
samala: samara, to which the author remarked 'in our opinion the correct
version is samala' (*^JI Ujula ^yxJ\). 6

3 e.g. Goldziher, Richtungen, 30.

4 Pace Goldziher, op. cit. 34; id., Studien ii, 201, 242-3; and Blachere, Histoire iii,
798 n. 1 ; see above, pp. 176-7, I pp. 51-2.

5 Goldziher, Zdhiriten, 33—4: on consensus (ijma).

6 Kitab al-nasikh taal-mansukh, MS Ahmet III, 143, 95 r .

1 Iggeret Tertian, 80 lines 15-16.

2 Jassas, Ahkdm al-Qur'dn iii, 28-9.

PRINCIPLES OF EXEGESIS

\ For Ibn 'Arabi the problems relating to transmission of Sunna were
clearly of polemical character. 1 His point of departure was the familiar
charge against the Jews that they had altered God's word, e.g. in Q. 2: 59

p^J J-J ^jJI jj. Vji' I _^JlU ^jjJ! JjLi, referring specifically to hitta in the
preceding verse (and Q. 7: 161), a much-disputed term thought to exhibit
the malicious alteration of nson (sin) into nun (wheat; Arabic hinta). 2
Whatever the historical background to that charge may have been, it was
interpreted in the Muslim exegetical tradition as proving wilful distortion
of a divine command. Ibn 'Arabi's argument was that if an act of worship
(ta'abbud) depended upon the actual expression (lafz) employed, there
could then be no question of paraphrastic transmission. The latter was in
any case an indulgence limited to the companions of the prophet, and only
for non-liturgical formulations. Now, Q. 2: 59 (like 7: 162) is one of many
Quranic passages assumed by exegetes to refer to a conscious and malicious
distortion of the word of God. 3 The three technical terms tabdil, tahrif, and
hitman, employed to describe that procedure are amply attested in scrip-
ture, each in a variety of contexts permitting association of the act with
written texts, and thus tantamount to forgery (e.g. Q. 10: 15, 4: 46, 2: 174,
respectively). Haggadic embellishment of the charge turned mostly upon
the absence from Hebrew scripture of proof-texts announcing the mission
of Muhammad, symbolized by the many stories connected with the
Jewish convert to Islam Ka'b al-Ahbar. 4 No one in Medina was more
familiar with the Torah than Ka'b, and in one account he specified ten
Quranic verses (six of them concerned Abraham) which the Jews had
allegedly erased from their scripture because they contained predictions of

the advent of Islam: U> jj > ^ ^ U^tjj (l^udL?) l$^. tl Ak s

For the halakhists recourse to tactics such as those was hardly adequate,
and the accusation of forgery was gradually elaborated to include both
textual alteration and exegetical error, the latter in the sense of intention-
ally false interpretation. Intermediate and combined positions were also
possible, and each had its origin in Judaeo-Muslim polemic. 6 The implica-
tions for Quranic exegesis were articulated by Jassas: the concept tahrif in

Q. 5 : 13 <u^lj^ (J^l <3y>*i was limited to interpretation (ta'tnl) of the
kind which necessarily resulted from lack of or from arbitrary method,

1 Akkdm al-Qur'dn i, 10.

2 See Hirschfeld, Researches, 107; Speyer, Erzahlungen, 337~8; and the discussion in
Paret, Der Koran, 1 9-20 ad loc.

3 See above, pp. 154-5; H PP- 6 3> 76.

4 See above, II pp. 75-6; and Rabin, Qumran, 116, 118, 123-4.

5 Perlmann, 'Another Ka'b al-Ahbar story', 48-58.

* See Steinschneider, Poletmsche Literatur, 320-2, 392; GoWziher, 'Polemik', 345.
364-72; Schreiner, 'Zur Geschichte', 599-601, 613-14, 626-8, 634-5, 64°-^; Hirschfeld,
'Mohammedan criticism', 222-40.

QURANIC STUDIES

e.g. elucidation of the mutashabihat without reference to the muhkamat.
Tahrif in the specific sense of alteration (taghyirjtaghayyur) could not be
applied to any text whose attestation and transmission were derived from
widespread authority (istifddajtawdtur), since that would diminish and
even cancel the value of all tradition. 1 The semantic position thus defined
for tahrif was rather more constructive than one from which merely
abusive assaults on the integrity of Hebrew scripture could be made and,
not unexpectedly, generated a more sophisticated polemic. For charges of
maliciously false interpretation ( !) levelled at one another by communities
sharing a set of scriptures were a commonplace of sectarian strife. 2 That
the quality of invective elaborated there might serve as model for dispute
about the content of another scripture seems not unreasonable: the
language of polemic is remarkably uniform. But in a primarily consonantal
text interpretation might easily involve argument about variants in the
received tradition, for which of course tahrif as textual 'alteration' could
prove a convenient tag. It will, however, be useful to remember that the
existence of textual variants presupposed rather than prefigured divergent
interpretations. 3

The actual condition of the textus receptus could always be justified, and
even turned to advantage. For example, absence of die basmala at the
beginning of Surat al-Baraa caused Ibn 'Arabi to observe not only that
the unity of subject-matter in Suras 8 and 9 had naturally (and logically)
precluded insertion of the formula, but also that such was proof of the
divine origin of analogy. 4 His argument was that the similarity (tashbth) of
the two suras had, in the absence of specific textual indication ('tnda 'adm
al-nass), caused the companions of the prophet ( !) to resort to analogical
juxtaposition (qiyds al-shabah) of the two originally separate revelations.
That solution is less far-fetched than Noldeke-Schwally appeared to
believe, as ought to be clear from the use made of verses from both suras
in discussion of the prescriptions for Holy War. 5 Not only analogy, but all
other methods of demonstrative proof (sd'ir durub al-istidldl) could, in the
opinion of Jassas, be derived from the text of scripture. 6 Designation of the
book as 'clarification of all things' (tibydn K-kull shay') constituted an
invitation to the exercise of logic, permissible in the absence of explicit

answers in Qur'an and Sunna, or of consensus: Aj^Uai Ju*J .J lil

1 Ahkdm al-Qur'dn ii, 398-9.

2 For Qumran and the Karaites, see Wieder, Scrolls, 135-53, 161-3; and cf. e.g. Jere-
miah 23 : 36.

3 See below, pp. 202-8.

4 Ahkdm al-Qur'dn i, 366 ad loc.; cf. GdQ ii, 79-81.

5 e.g. in Tabari, Kitdb Ikhtildf al-fuqahd' , passim.

6 Ahkdm al-Qur'dn iii, 189-90 ad Q. 16: 89.

PRINCIPLES OF EXEGESIS 191

For Jassas consensus represented the uninterrupted transmission of com-
munity opinion from the time of the prophet, the term umma (nation) in
Q. 2: 143 being interpreted as a general ('dmm) not a specific (khdss)
epithet for 'witness'. 1

Instances of incompatibility or of conflict between scriptural passages
containing juridical material were resolved by the halakhists with the aid of
three related but distinct hermeneutical devices: takhsis (specification),
tafsir (here corroboration), and naskh (abrogation). Takhsis provided a
means of linking a general statement to a particular situation, and was
generously employed in identifying the rn.ushrih.un of Q. 9: 5 and the
muharibun of Q. 5: 33.2 Application of takhsis presupposed a very flexible
standard of generality against which particularity was measured: it was
thus found that the adjective kull (all/each) as well as the various relative
pronouns, the definite article, and even indefinite predication, could be
both general and particular. 3 The option in any given instance entailed
almost invariably acknowledgement or rejection of a point of doctrine
hardly adumbrated in the scriptural passage itself. Halakhic takhsis, in
brief, depended upon the kind of analogy called ta'lil, or inference from a
tertium comparationis.* An example was the 'specification' of Q. 5 : 3 (pro-
hibition of carrion) by Q. 5 : 96 (extension to carrion from the sea) and by
6: 145 (extension to flowing, as opposed to coagulated blood), in which the
ratio ('ilia) was contained in the opposition haram : haldl. 5 Another case was
specification of Q. 24: 2 (punishment for fornication) by Q. 4: 25 (extension
to the betrothed), by reference to the (general) fact of punishment. 6 More
often than not it was Sunna to which appeal was made for specification of
Qur'an, a procedure defended by both Shafi'i and Jassas, and of which an
example was specification of Q. 2: 275 (prohibition of usury) by traditions
extending the sanction to similar transactions, e.g. 'araya. 1 Exclusion of the
slave from inheritance by application of takhsis to Q. 4: 11-12 and 2: 180
belongs to the same category, though there the procedure was probably a
device to conceal the priority of an established legal maxim. 8

Like takhsis, the technical term tafsir in halakhic usage exhibits a for-
mula of harmonization designed to restrict the sphere of abrogation (naskh),
itself the final court of appeal in the more general effort to demonstrate a
scriptural source of authority for all Islamic law. An essential advantage of
both takhsis and tafsir, and to a considerable extent the difference between
them and naskh, was that Quranic verses so treated remained effective

1 Ahkdm al-Qur'dn i, 88-90 ad loc. * See above, pp. 185-7.

3 Cf. SuyutI, Itqdn, nam' 45: iii, 43-51, esp. 43-5. 4 See above, pp. 166-8.

5 Suyuti, Itqdn iii, 44, 47. 6 Itqdn iii, 48.

7 Itqdn, loc. cit. ; see above, pp. 175, 188.

8 Pace SuyutI, Itqdn iii, 48, who considered it specification by consensus ; see Schacht,
Origins, 184-5.

192

QURANIC STUDIES

(muhkamat). An example of tafsir, though the term was not used, was the
argument of Nahhas about the relevance of Q. 9 : 5 {ay at al-sayf) and 47: 4
(recommending leniency) to the treatment of prisoners of war: the two
verses were not mutually exclusive but corroborative, and the option lay
with the imam. 1 The secondary role of scripture in that argument is,
incidentally, illustrated by its appeal to prophetical traditions pertinent to
the conquest of Mecca. Similar reasoning was adduced by Abu 'Ubayd, on
the authority of Ibn 'Abbas, in a discussion of the lex talionis (qisds), for
which Q. 5: 45 was held by some to abrogate the stricter ruling of 2 : 178. 2
With regard, however, to the distinction between free men and slaves in
Q. 2 : 178, Abu 'Ubayd argued that the locution 'a life for a life' in 5 : 45 was
not the abrogant but, rather, the corroboration of equality within the

separate categories of retaliation: ajJldl J ^ydl i>Vl J ijl <£j LJ

Lgj^Ai (j "^^li j^^i ti i^*«^J

OjiJI (j f jJi 'ij+Jud^ «JjUJI (j ^J$\ <jl t£lj Ail oLSitv U.a--^ l»4Jl5sJj
p g L-i SjjL»ju jt jf-VI (jl jj* _j* tj^k j*^" aJjj (jl Jjtxj

Ju-jJI jji. The use of ta'azvwala in this passage for 'interpretation' might

be thought to confirm the specialized meaning of mufassira as 'corrobora-
tion', antithetically juxtaposed to nasikha (abrogant).

That the hermeneutical principle of abrogation (naskh) did not refer to
supersession of earlier divinely revealed statutes (shara'i' al-anbiya') by
Muhammad's law {shari'at muhammad) was expressly articulated by
Jassas. 3 It was, rather, the instrument by means of which particular state-
ments (khass) could be distinguished from general ones ('dmm), and
polyvalent utterances (mutashdbik) referred to univalent ones (muhkam). In
that formulation takhsis (specification) was subsumed under the general
rubric naskh, but though certain of the phenomena which regularly
appeared in discussions of scriptural abrogation were sometimes desig-
nated takhsis, it is both convenient and realistic to distinguish the two. The
concept of Islam as the supersession of earlier dispensations was of course
not alien to the Muslim exegetical tradition, and might even derive some

support from scripture, namely Q. 5: 50 jjjlo 2LUUJI ^>S\

dj^ji L£a- «U|| t y t . Explicit reference to a 'pagan dispensation' (kukm

al-jdhiliyya) was interpreted by Goitein as signalling the commencement
of a specifically Muhammadan legislation derived from material up to that
point diffuse and only paraenetically expressed. 4 Acceptance of Jahiliyya
as a temporal rather than psychological or sociological concept and
acquiescence in the utility of dating the contents of the Qur'an reflect a

1 See above, pp. 184-5.

1 Kitab al-nasikh wal-mansukh, MS Ahmet HI, 143, 93 r-v .

3 AJikam d-Qur'an i, 58-60. * Goitein, 'Birth-hour', 132-3.

PRINCIPLES OF EXEGESIS

193

view of Islamic origins which can hardly profit from modification of
details. The context of the entire passage Q. 5 : 42-51 is polemic and was so
acknowledged by Muslim exegetes, whose view of the prophet's juris-
diction can only with difficulty be related to pagan practice in the process
of replacement by divine decree. The historical circumstances envisaged
by, and if not invented then certainly elaborated within, the exegetical
literature were those surrounding Muhammad's celebrated confrontation
in foro externa with the rabbis of Medina. The essential truth elicited from
that story was, it may be recalled, that the role of the Muslim prophet lay
not in abrogation or supersession of the Mosaic Law ( !), but in its restor-
ation and fulfilment. 1

Now, in his monograph on the phenomena of abrogation, which apart
from the Risala of Shafi'i must be the earliest treatment of that subject, 2
Abu 'Ubayd found it convenient to link discussion of the penalty for
fornication with that of Muslim jurisdiction among non-Muslims. 3 In the
first section the author-argued that both Q. 4: 15-16 and 65: 1 (on those
justly charged with fornication) had been repealed by 24: 2 (which specified
the punishment as 100 lashes) and the stoning penalty. In one of the two
traditions adduced, both traced to Ibn 'Abbas, stoning was reserved for
those offenders who were muhsan (i.e. chaste, betrothed, possibly married,
free, Muslim, etc.). 4 Though the stoning part of the penalty was there
specified as Sunna, two versions of a tradition from 'Ubada b. Samit
immediately following leave no doubt that, Sunna or not, the stoning

penalty had been revealed to the prophet: J*». J5 ^js. Ijli. <U!l Jj*.j Jli
pr^-jij •UjcJ t— -r^lj -^ijSCJI Vs^'j J^r"^ ^sr 1 * < *^' -5

Thereafter, the author turned his attention to the problem of punishments
(hudud) to be inflicted upon dhimmts, and the question of the abrogation of
Q. 5 : 42 by 5 : 48. That the particular case envisaged was fornication is
clear from the gloss of bi-md anzala Hlah in Q. 5 : 48 as al-rajm (stoning),
thus characterized again as 'revelation', though that point was followed by
discussions of the lex talionis and the muhdribUn. 6 In the second section

the abrogation of Q. 5 : 42 ^s. \J> j£.\ j\ j^-a j^C^U iijjU by 5 : 48
«U)I Jjil Uj |t a-j p5C».li was also illustrated by the stoning penalty, this
time specified as the prophet's stoning of the two Jews <£Sj$J\

1 See above, II pp. 70-1. 2 See above, p. 175.

3 Kitab al-nasikh vial-mansukh, MS Ahmet III, 143 : 'Bab al-hudtid', 88 T -o7 r , and
'Al-IJukm bayna ahl al-dhimma', i72 r -4 r .
+ Cf. Ibn Qutayba, Ta'tvil, 391.

5 Abu 'Ubayd, op. cit. 8<) v and oo r : variant prefaced by a naturalistic description of the
prophet in a state of 'reception', widely employed in the hadith literature, e.g. Bukhara,
Sahih, 'Kitab al-Shahadat' 2, 15; and references Wensinck, Handbook, 162-3.

6 Abu 'Ubayd, 9o v -i r ; see above, pp. 192 and 186, respectively.
4389C75 H

194

QURANIC STUDIES

io j^JI j. 1 In the opinion of the Iraqi jurists that incident had established
the precedent for dealing with litigations among dhimmis, while the Hijazis
argued that it had taken place before imposition of poll-tax (jizya) and thus
did not afford a precedent for the later period, when administrative pro-
cedure tended towards judicial autonomy within dhimmi communities. Abu
'Ubayd rejected the latter position on two grounds: first, that traditions
did not in fact state that the stoning incident had taken place before the im-
position of jizya ; second, that even had that been so it would not preclude
dealing with litigation among dhimmis after imposition of the tax. Prior to
that condition in fact, they would not have been dhimmis but merely in
treaty relationship (hudnajmuzodda'a) with Islam like other infidel nations
{umatn al-shirk). The question of jurisdiction would not even have arisen. 2
But once established, the contract between the Muslim community and the
newly recognized dhimmis enabled the latter to resort to Islamic juris-
diction. The implication was that Q. 5 : 48 abrogated not only the circum-
stances described in 5 : 42, but also those of 5 : 50. 3 In the light of that
argument it must, I think, be conceded that hukm al-jdhiliyya (Q. 5 : 50)
referred to, or at least included, Jewish practice prior to the prophet's
intervention in that much-disputed litigation.

Though Abu 'Ubayd did not specify dyat al-rajm (stoning 'verse'), it is
clear from his presentation of the traditions from 'Ubada b. Samit that the
stoning 'penalty' was of revealed status, that is Sunna but not Qur'an. The
movement exhibited in the transposition of stoning penalty into stoning
verse 4 may be understood to reflect elevation of the Quranic text to canoni-
cal status : a source of legislative authority. The role of ay at al-rajm was
henceforth (beginning of the third/ninth century) intimately connected
not merely with the historical description of Judaeo-Muslim polemic, but
also with the principle of legislative repeal in scripture. As a piece of
historical evidence dyat al-rajm may be assessed by its haggadic projection,
and especially within the narrative framework of the Muhammadan
evangelium. As witness to a hermeneutical principle, it may be judged by its
halakhic value in establishing, in the context of disputatio fori, that scrip-
ture was the ultimate source of all legislation.

The literary forms generated by Muhammad's confrontation with the
rabbis of Medina were signalled by Goldziher in a study of Jewish practices
as described in Muslim literature. 5 That those represent variations upon
the archetypal theme informing the account of Jesus with the Pharisees
was not mentioned by Goldziher. The conclusion of Vajda that there must
be a nucleus of historical truth in the story cannot of course be ruled out,

1 Abu 'Ubayd, i7z r -4 r , esp. i73 r . 2 Abu 'Ubayd, I73 T -

3 Abu 'Ubayd, I74 r : sukt is glossed 'bribe' (rishwa).

* Variants of the latter were assembled by Suyuti, Itqdn iii, 72-3, 75-7; cf. GdQ i,
248-52. s See above, II p. 71 : reference n. 1.

PRINCIPLES OF EXEGESIS

1 95

S but neither can it be demonstrated. 1 The entire literary complex belongs to
a cycle of 'tests of true prophethood' widely distributed in Muslim exe-
getical literature, and hardly susceptible of positivist interpretation, such
as Muhammad's being forced by circumstances to abandon a harsher
penalty for adultery in favour of Rabbinic leniency, or as posterior justi-
fication of 'Umar's conduct in stoning for adultery. 2 In his impressive, if
not entirely convincing, mise en scene of the adulteress pericope (John 7 :
53-8: 11) Derrett stressed the qualification of witnesses, rather than the
nature of the punishment or its scriptural sanction, The assumption must
be that Jesus was approached not as a prophet (like Muhammad), but as a
rabbi especially competent in the laws of evidence. Such was undoubtedly
the significance of a link between the pericope and the story of Susanna and
the elders, but not, I think, of its inclusion in the Muhammadan evangelium. 3
For the slightly modified Muslim version, the observation of Ibn al-
Jawzi that the 'Torah passage' adduced by the rabbis contained the
(Quranic) stipulation on the necessity of four ( !) witnesses to the aet^of
fornication (Q. 24: 4), might be thought sufficient indication of the
polemical purpose for which the account of Muhammad's triumph had
been devised, namely, the maliciously concealed coincidence of Muslim
and Jewish scripture: ViT K»D iWS nVbT mm nniK in» NiK

Din rfrnsaVN VaVx. 4 Reference in the same report to Jewish abandon-
ment of the scriptural (hadd) penalties after destruction of the Second
Temple (idh kan al-mulk land) served as motive to Muhammad's restora-
tion (ihya) of the Mosaic Law, and may be exegetically related to hukm
al-jahiliyya in Q. 5 : 50.

Halakhic elaboration of the story had as point of departure the interest-
ing circumstance that neither stoning penalty nor stoning verse was in-
cluded in the canon. That such gave less cause for alarm than might be
supposed will be clear if it is understood that the principle of abrogation,
as well as the development of ahkam (halakhot), concerned the entire, very
flexible, corpus of revelation, of which the Qur'an was only part. Efforts to
relate the phenomena of abrogation to the canonical text exhibit not a
necessity but merely a tendency to seek scriptural support for positions
already occupied and for the most part consolidated. Neither the principle
of abrogation nor the formulation of ahkam required support in scripture
until scripture itself came into being as a result of external pressure in
polemic. For ay at al-rajm Burton put the case well: 'The "process" here
promised was later "appointed" in the revelation of the stoning penalty.
The stoning penalty, and not the Qur'an, was thus the historical source of

1 Vajda, 'Juifs et Musulmans', 93-9.

2 See Vajda, loc. cit.: Hirschfeld, Researches, 103, 137.

3 Law in the New Testament, 156-88.

* Goldziher, 'Usages juifs', 326: for the description of consummation cf. Talm. Babl.
Makkot 7a.

i 9 6 QURANIC STUDIES

the "stoning verse".' 1 'Later appointed' is a reference to a variant reading
in Q. 2: 106 l^Li* jl L^u j-jej oii (UL^jj) lg-*Ji jl AjT ,y» ^<*jJ U,
which became the locus classicus for the doctrine of abrogation. The
variant itself, 'we defer' as opposed to 'we cause to forget', drew upon
that category of asbdb al-nuzul which included the concepts of promise
and ratification, by whose means the arbitrary data on Quranic chronology
could be conveniently neutralized. 2 For Jassas, 3 a variant reading was un-
necessary: the verb ansa might signify either forgetting (nisyany or defer-
ment (ta'khir), both in the sense of exchange (tabdil) for the public weal
{maslaha: ius propter utilitatem publicam). 5 Examples were the exchange of
qibla (Mecca for Jerusalem) and the repeal of Q. 8 : 65 by 8 : 66 (takhfif) , which
together illustrated the principle that Qur'an could abrogate Sunna and
Qur'an, the term ay a in Q. 2 : 1 06 being a reference not to the ipsissima verba
of scripture (tildwa) but to the precept contained or implied therein (hukm).
That particular view of the range of abrogation was only one of five, set out
by Nahhas: 6 Qur'an could abrogate Qur'an and Sunna (the argument of
the Kufans); Qur'an, but not Surma, could abrogate Qur'an (of Shafi'I,
whose concern with Qur'an was anyway peripheral) ; 7 Sunna could abro-
gate Qur'an and Sunna (anonymous); Sunna, but not Qur'an, could
abrogate Sunna (also anonymous); these categories were not mutually
exclusive and each case had to be judged on its merits (ascribed to Muham-
mad b. Shuja'). In practice only the last could survive as a working prin-
ciple, and few if any cases were ever decided by an actual appeal to the
priority of one over another kind of revelation.

Of four major works devoted to the phenomena of abrogation, two were
concerned primarily to demonstrate the presence of such in the text of
scripture: those of Nahhas and of Hibatallah. Each introduced his work
with the story of 'All b. Abi T^lib and the preacher in Kufa who was
banished from the mosque for not knowing his principles of abrogation. 8
That level of discourse was hardly modified throughout the treatise of

Hibatallah, who defined naskh as removal/cancellation (^ij yt}\ ^S' l ^

1 Burton, 'Cranes', 261. 2 See above, pp. 175-7.

3 Ahkam al-Qur'dn i, 58-60.

4 On the problems provoked by the interpretation nisydn, see Burton, 'Cranes', 260-3 ;
to which might be added the observation that in the Quranic lexicon ansa is not infrequently
connected with satanic agency (e.g. Q. 6: 68, 12: 42, 18: 63, 58: 19), and may have
been employed as metaphorical counterpoint to the verb alqd (e.g. Q. 4: 171). The
conjecture is in no way weakened by explicit reference to divine agency in Q. 2: 106
and 59: 1 9> and to satanic agency with alqd in 20: 87 and 22: 52-3, and could be related
to the neutral concept 'inspiration' by mediation, characteristic of Muslim prophetology;
see above, II pp. 58-61.

5 See Goldziher, 'IstishaV, 229-30.

6 NabJjas, Kitdb al-ndsikh wal-mansiikh, 6-7.

7 Cf. Schacht, Origins, 15 : tacit correction of Goldziher, Studien ii, 20.

8 Nahhas, op. cit. 5-6; Hibatallah, Kitdb al-ndsikh wal-mansukh, 3-5.

PRINCIPLES OF EXEGESIS i 97

^LII) and limited its incidence in the Qur'an to explicit command and pro-
hibition (amr zoa-nahy) or to reports (akkbar) containing or implying such.
Of these he found 239 instances in 71 suras, a consequence of massive and
undifferentiated assertion, rather than of subtle and reasoned analysis : 107
occurred in Suras 2-9 and Q. 9: 5 (ayat al-sayf) figured as abrogant 124
times. 1 For Nahhas the problem was less simple. The term naskh might
mean cessation (izala) or transfer (naql), and the Quranic principle of
abrogation was based upon the latter. 2 He also found expedient a distinc-
tion between naskh and bada' : the former might apply only to command
and prohibition, the latter to instances of contradiction or inconsistency
(apparent !) in reports, to which one could attach a limitation in terms of
altered time or place, e.g. the changing circumstances of narrative, as in
the stories of the prophets. 3 Of abrogation according to his own definition
Nahhas found 137 instances in 48 was, of which 75 appeared in Suras 2-9.

Common to Hibatallah and Nahhas was a typology of the modes of
abrogation attested in scripture. 4 These were three (the authors employed
different terminology, and Nahhas, perhaps for the sake of lexical tidiness,
included a fourth which identified naskh in the sense of 'copy') ; abrogation

of both wording and ruling (oj}*dlj |*5oeJI ^J/<u5Ca-j -dti.
abrogation of wording but not of ruling (^j/i^a. aj^
p£>JI dji Sy/dl); abrogation of ruling but not of wording (p5oJI
•vki. Jhj ^uiC» ^Jjej^cil Oj->)- That these formal distinctions con-
tained, and were also very likely meant to conceal, essentially irreconcil-
able views on the constituent parts of Muslim law was indicated by Burton. 5
For my purpose here it is sufficient to state that the modes of abrogation
set out by Nahhas and Hibatallah reflect a concerted effort to identify naskh
as an originally Quranic phenomenon. Once seen to enjoy scriptural
sanction, the principle of abrogation could in theory and with impunity be
applied across the entire range of source materials for the formulation of
Muslim law.

A favourite example of naskh in the Quranic text was the alleged repeal
of Q. 8 : 65 by 8 : 66, where the number of enemy which Muslims were
expected successfully to oppose in combat was reduced (takhftf: lightened)

fromaratio of 10: 1 to 2: 1, e.g. inTabari: 3u l^t^ji p-fr^ (Juiei

gjlri]. 6 Belonging to the third mode (above), that instance of abrogation

was typical of nearly all those adduced by Nahhas and Hibatallah, pride of

1 See above, pp. 183-5. 2 Nahhas, 8.

3 Nahhas, 10-11; see Goldziher, EI, s.v. Bada'.

4 Hibatallah, 5-6 ; Nahhas, 8-9.

5 'Cranes', 358-64, and discussed in detail in the study referred to there, 249 n. 4.

6 Tqfs&r x, 27 ad loc.

198

QURANIC STUDIES

place going of course to dyat al-sayf. To illustrate the first mode (above),
the much-discussed 'satanic verses' alleged originally to have been at Q. 53 :
19-22 and abrogated by 22: 52 were invariably adduced, and the several
motives which led to that assertion were analysed at length in Burton's
study. 1 It was to demonstrate the second mode (above) of abrogation that
dyat al-rajm was commonly introduced into halakhic controversy, namely,
as an example of a valid ruling whose wording had been removed from
the text of scripture. Now, that the origin of the stoning verse lay in the
account of the stoning penalty has been proposed. That the origin of the
stoning penalty may be sought in the haggadic topoi traditionally employed
to illustrate the test(s) of 'true prophethood' seems more than likely.
Juridical appropriation of that particular topos, however transparent the
motive, was the inevitable consequence of pressure compelling recognition
of the Qur'an as a source of legislation equal to the Sunna. The adjustment
exhibited in the transposition: stoning penalty -^stoning jyerse was only
necessary after establishment of the canonical text of revelation. Accep-
tance of the transposition, or even of the second mode of abrogation, was
not universal: Nahhas, for example, recognized the isnadoi dyat al-rajm as
sound but insisted upon regarding it as Sunna, and thus not as evidence of
Quranic abrogation. 2

Two other works dealing with naskh were less concerned with the speci-
fically Quranic data and rather more with the principle of abrogation as a
valid juridical premiss. From what has been said of Abu ' Ubayd's treatise
on ndsikh wa-mansukh it ought to be clear that status as Qur'an or Sunna
was hardly operative in his formulation of the rules. Arrangement of the
material is topical, rather than by Quranic division, and most, if not quite
all, chapters bear the subtitle 'that which abrogates and is abrogated in
both Qur'an (kitdb) and Sunna'. The twenty-seven chapters contain the
conventional range of ahkdm, e.g. Prayer (fols. 8 v -i3 v ), Almsgiving (fols.
r4 r -ia v ), Fasting (fols. 2o T -^6 T ), Marriage (fols. 46 r -74 r ), Divorce (fols.
74 r -88 r ), etc. 3 "Whether Abu 'Ubayd (d. 223/838) was the first scholar to
treat monographically ( !) the subject of abrogation can probably best be
answered with reference to the chronological development of Quranic
studies, rather than to reports of earlier written works. One such is the
ascription to Zuhri (d. 124/742) of a book entitled Ndsikh wa-mansukh fil-
Qur'dn.* Like the quite worthless Tanzil al-Qur'dn attributed to the same
author, 5 the work on abrogation is preserved in the recension of the Sufi

1 Burton, 'Cranes': to the 'historical' (Orientalist) references mentioned there, 246-9,
may be added Andrae, Person, 129-33.

2 Kitab al-nasikh uial-mansukh, 9; cf. the dissenting opinions recorded in SuyutI,
Itqdn iii, 72-^7; and the anonymous ikhtildf in Ash'ari, Maqdldt, 607-11.

3 Kitdb al-nasikh voal-mansukh, MS Ahmet III 143 ; cf. the arrangement in Muqa-
til's halakhic treatise, MS BM Or. 6333, above pp. 173-4.

4 GAS i, 283 no. 4; MS Beyazit 445. 5 gee above, pp. 179-80.

PRINCIPLES OF EXEGESIS

199

exegete Sulami (d. 412/1021) and will, even if it exists, hardly contribute to
the history of naskh theories prior to Abu 'Ubayd. However that may be,
the manuscript Beyazit445 is not the work of Zuhri-Sulaml, but of 'Abdal-
qahir Baghdad! (d. 429/1038).' The significance of Baghdadi's position in
the study of naskh is that long after Nahhas (d. 338/950) and some time
after Hibatallah (d. 410/1019) a work methodologically similar to that of
Abu 'Ubayd should be written at all. The author's primary concern was
with the theoretical elaboration of naskh, its justification as a juridical
principle, and its attestation in Sunna as well as in Quran. In the three
chapters containing instances of Quranic abrogation the matter was intro-
duced in the form of dispute (ikhtilaf) and consensus {ijma'), and the total
number of verses adduced only 59. 2 In an introductory chapter the notion
of naskh as removal/cancellation (raf) or cessation (izdla) was rejected in
favour of a combination of specification (takhsis) and transfer (tahwil). 3

The author's conclusion may be cited: 4 JJi ljj_<L> ^~JI J I* ^ >_?

AZxJjZj U.J Vl ^j-i (j l&) dUwa^JI <*-»jJj ksOjLi Aij^J

L'JUfr ^e^aJI yt. Il&j <u)l Jl»o. That naskh might indeed refer to

the abrogation by Muhammad's revelation of an earlier divinely revealed
statute (here the 'law of Abraham') was a possibility never quite sup-
pressed. 5

Abrogation as supersession of earlier dispensations was of course funda-
mental to the character of Judaeo-Christian polemic. That the Jews
allegedly rejected the specifically Islamic principle of naskh as in their
opinion nothing more than retraction/substitution owing to the emergence
of new circumstances (badd') was one among a number of problems raised
by SuyutI, himself satisfied that God was capable of reversing ('aks) any of
His actions or decisions, and that such had indeed been many times
attested in the history of divine revelation. 6 The allegation is puzzling,
since the retraction, reversal, and change of God's word was familiar
enough from Hebrew scripture, e.g. 2 Kings 9: 1-12 and Hosea 1 : 4-5, as
well as the crucial 'new covenant' of Jeremiah 31: 31. 7 Moreover, the
integrity of the Mosaic Law, explicitly stated in Deuteronomy 13: 1 (and
cf. Qohelet 3 : 14), was never intended to preclude progressive modifi-
cation according to altered circumstances in the community. The necessity
of and capacity for modification is amply attested in the terminology of

1 GAL i, 385 ; GdQ i, 54 n. 1, ii, 16 n. 5: MS Petermann I, 555.

2 Baghdad!, MS Beyazit 445, 7 r ~46 r , 46 r -7i r , 7i v -4 v -

3 Baghdad!, 2 r -3 r .
* Baghdad!, j6*.

5 See above, pp. 192-3; and Hasan, 'Theory of naskh', esp. 182-3, where naskh as
abrogation was denied, but acknowledged as the supersession of earlier revelation(s).

6 Itqdn iii, 6; see above, p. 197.

7 Cf. Eissfeldt, Einleitung, 694; and above, I pp. 11-12.

200

QURANIC STUDIES

Rabbinic exegesis, e.g. in the formula 1 rpVn Voaa sma, the antithesis 2
0*7 vs Vtra or "EC, and the notion of normative, as contrasted with
prescriptive, legislation contained in the terms 3 WJB and flK "P" 7 -
The extent to which those formulations may be interpreted as evi-
dence of abrogation depends of course upon the precise meaning of
that term. It must by now be clear that if there was ever agreement among
Muslim scholars on the semantic content of naskh, such would indicate
general acceptance of 'change' or 'transfer', reflected in the terms naql and
tahwil. The genuinely halakhic employment of naskh meant the un-
ceasing interpretation of scripture by reference to the example of the
prophet, and for the halakhists both sources were equally part of revela-
tion. Discussions of specifically Quranic abrogation, on the other hand,
represented a polemical defence of the text of scripture, and were only
marginally related to the formulation of law.

And yet it was precisely that latter aspect of the Muslim doctrine of
abrogation which informed Judaeo-Muslim controversy about its exe-
getical relevance. For example, Saadya's rejection of naskh consisted en-
tirely of arguments designed to prove that there were no contradictions in
the Biblical text, but only occasional passages which could be seen to
require hermeneutical complement or specification, e.g. whether sacrifice
or circumcision could be performed on the Sabbath. 4 Similarly, the five
points adduced and rejected as evidence of naskh by Sa'd b. Mansur (Ibn
Kammuna) reflect an attitude towards the text of scripture nearly un-
related to the eventual necessity of halakhic modification of its contents. 5
In his rebuttal, on the other hand, the author invoked the traditional
arsenal of Rabbinic terminology to prove that such modification was

possible, e.g. JjS c^: Vlp ro. 6 Those arguments for naskh in Hebrew

scripture adduced and so easily dismissed by Jewish apologists can hardly
be said to exhibit either the subtlety or the range of Muslim discussions of
abrogation. A degree of misrepresentation or, at least, of incomplete
representation of opposing views is not unexpected in polemic, but that
the methodological proximity, even identity, of Rabbinic and Sunn!
Muslim halakhah should be ignored, or suppressed, in the polemical
literature might be thought to require an explanation. Among Jewish
scholars were some who did in fact acknowledge the presence of naskh in

1 Bacher, Terminologie ii, 53-6; cf. Zunz, Vortrage, 54 n. (e); and Elbogen, Gottes-
dienst, 356.

2 Bacher, Terminologie i, 170-2, ii, 142, 186-9, 239; Gerhardsson, Memory, 07 n. 7,
233 n. 2, 264, 287.

3 Bacher, op. cit. i, 25, ii, 40-1 ; Gerhardsson, op. cit., 117, 182, 256, 305, 317-20.

4 Saadya, Kitab al-Amdndt, 128-45, esp. 140-5 ; cf. Schreiner, 'Zur Geschichte', 604-6.

5 Ibn Kammuna, Tancph al-dbfidth, 45-7; cf. Steinschneider, Polemische Literatur,
324-

6 Tanqih al-abfidth, 47.

PRINCIPLES OF EXEGESIS

201

Hebrew scripture, for example, Abraham b. David, 1 who regarded as
permissible abrogation in the sense of recognizing the temporal and/or
spatial limits of certain divine prescriptions, in other words badd* or
takksis. But the preoccupation of Jewish apologists was naturally, in the
light of Muslim claims that predictions of the advent of Islam had been
removed from Hebrew scripture, with the unassailable character of their
book, and one cannot help but suspect that it was abrogation in the sense of
supersession which, designated naskh in the polemical literature, was being
unconditionally rejected.

That was not the meaning of naskh in Muslim halakhic exegesis, but the
overtones of abrogation as reference to the supersession of earlier revela-
tions had never been quite eliminated, e.g. in Baghdad!, and could even be
derived from the principal Quranic loci probantes traditionally adduced in
support of the doctrine: Q. 2: 106 but also 22: 52, 16: 101, 17: 76, and
and 13: 39. As a contribution to the history of terminological complexity,
the lapsus calami in the Hebrew translation of Mairnonides' Pereq Heleq,
which resulted in rendering "|0a not as VlD3 (naskh) but as j?rtS? (naql), is
not without interest: it was the abrogation of the Mosaic Law by subse-
quent revelation(s) which the author was concerned to deny. 2 One can
hardly insist that the course of polemic on the subject of abrogation was
influenced exclusively by exploitation, or ignorance, of the ambiguity
inherent in the Arabic term naskh. It does, however, seem clear that the
real dispute was not about differences of exegetical procedure, of which
there were virtually none, but about the respective claims of Torah and
Qur'an to be the word of God. 3

In the formulation of ahkam, which could only retrospectively be under-
stood as the derivation of law from revelation, the fact of a canonical text of
scripture was, if not quite a hindrance, then of very little help and prob-
ably regarded as something of a challenge. No element in either the style or
the structure of halakhic exegesis points unmistakably to the necessity, or
even to the existence, of the canon as ultimately preserved and transmitted.
It may of course be argued that the rudiments at least of a comparative
method can be inferred from the chronological arrangement of scriptural
passages, from the juxtaposition of nmhkam and mutashabih, and from the
opposition of nasikh and mansQkh. But the comparison was of parts to parts,
not of parts to a whole. 4 As hermeneutical instruments chronology,

1 Cf. Steinschneider, op. cit., 353 no. 5 ; Schreiner, op. cit., 635-8.

2 See Hyman, 'Maimonides* "Thirteen Principles" *, 128 n. 58.

3 And thus more approximately related to the Muslim charges of forgery (tahrtf),
see above, pp. 1 89-90 ; on procedural similarity it may be noted that one of Ibn Kammuna's
arguments, Tanqih, 46, against naskh (sic) was that scripture commanded obedience to the
prescriptions articulated by successive prophets; see above, pp. 174-5.

* A halakhic midrash is unattested before Quitubi (d. 671/1272), cf. GAL i, 415-16,
Suppl. I, 737.

202

QURANIC STUDIES

analogy, and abrogation were, in respect of the existence of the canon,
neutral. Similarly, the employment of traditions (akadtth) to link Sunna
and Qur'an emphasized the role of revelation but not exclusively of scrip-
ture. The dichotomy of 'Qur'an as document' and 'Qur'an as source'
proposed by Burton, while not without a certain methodological utility, is
misleading if meant to postulate the historical existence of the canonical
text before it became a source of law. 1

Logically, it seems to me quite impossible that canonization should have
preceded, not succeeded, recognition of the authority of scripture within
the Muslim community. Chronologically, the data of Arabic literature
cannot be said to attest to the existence of the canon before the beginning of
the third/ninth century. These tentative and admittedly conjectural con-
clusions might be thought to derive some support from the form of scrip-
ture in halakhic argumentation: the practice of adducing selected and
discrete passages provides negative evidence of a kind comparable to the
absence of explicit reference to the Qur'an in other related contexts. 2
Moreover, the marginal character of lexical, grammatical, and syntactical
analysis in the work of the halakhists indicates little concern for a ne
varietur or even relatively stable text of scripture. 3 My own hypothesis,
that establishment of such a text presupposed rather than prefigured
acceptance of the Qur'an as a source of law, gains some strength from the
sudden efflorescence of masoretic exegesis soon after the literary formu-
lation of the ahkdm.

4. MASORETIC EXEGESIS

The Quranic masorah consists basically of three elements: lexical explana-
tion, grammatical analysis, and an agreed apparatus of variant readings.
Its elaboration required two exegetical instruments: textual analogy and
periphrasis, as well as the introduction of evidence from a large and con-
veniently flexible corpus of Arabic poetry. A single 'non-textual' com-
ponent of the masorah, and curiously, the only one adduced in the 'royal'
Egyptian edition of the Qur'an, is designation of the place of revelation
{mawatin al-nuzul), the purpose and derivation of which were not maso-
retic but halakhic. 4 But even that material is of some comparative, metho-
dological value for a study of the masorah proper, especially of the variae
lectiones. Ascription and transmission, both of information on the circum-
stances of revelation and of reports on variant readings, were formulated
exclusively as traditions from the companions of the prophet, and are thus
subject to the analytical criteria appropriate to, say, legal and 'historical'
traditions. It might be argued that a possible exception to this rule is

1 'Cranes', 251-2, 259. 2 See above, I pp. 44-5, II pp. 82-3.

3 See above, III pp. 100-1. « See above, pp. 177-81.

PRINCIPLES OF EXEGESIS

represented by the existence of 'regional' codices, but those appear to be
not only later than the 'companion' codices, but like them also largely
Active. 1 Of genuinely textual variants exhibiting material deviation from
the canonical text of revelation, such as are available for Hebrew and
Christian scripture, there are none. The Quranic masorah is in fact entirely
exegetical, 2 even where its contents have been transmitted in the guise of
textual variants.

An example will illustrate this not uninteresting phenomenon: in what
appears to be the earliest collection of variants (masdhif) from the con-
sonantal text of the 'Uthmanic recension (imam), a chapter entitled t_iU
[jTjDI J] i=?tJI l& ^yJ! oj/JI j-j J^ljjJI in the Fada'il al-

Qur'dn of Abu 'Ubayd (d. 224/838), Q. 18: 79 was rendered O&j
Lyai. S*JL<? iLuiLv JS* and the intrusive (non-canonical) 'sdliha'

attributed to the codex of Ubayy b. Ka'b. 3 Using the same source,
Bergstrasser-Pretzl described sdliha^sGund, in good repair) as an aetio-
logical addition (motivierender Zusatz), an obvious assessment from the
point of view at least of the canon's textual integrity. 4 But attribution to
Ubayy was arbitrary: the reading with sdliha appeared not only there
but also in the codices of 'Abdallah b. Mas'ud and Ibn 'Abbas. 5 The
transparency of that device, by means of which an exegetical gloss could
be construed as evidence of a textual variant, emerges from examination

of Muqatil's treatment of Q. 18: 79 Lyai- *>JL« azJu* JS - .i^L

Lji&t jy. Lyai- l>j*« j^Ju (7: 190) UJL<? L*li"T Lis <fbl?c~«, Jjjif'. 6

Were it not for the characteristically fluid boundaries between text and
commentary, 7 the epithet sdliha could justifiably be regarded as scriptural
and glossed by the analogous usage in Q. 7: 190. On the other hand, the
further epithets sahiha and savnyya (both signifying 'sound, in good repair')
could be interpreted as supercommentary, with which haggadic exegesis
was liberally strewn. 8 But whatever the textual state of Muqatil's scripture,
the process by which 'variants' to the 'Uthmanic canon were produced
and allocated to one or more of the 'companion' codices is worthy of
notice. The examples collected by Goldziher can hardly be interpreted
as other than exegesis to the canonical text. 9 Goldziher's understanding
of those phenomena as evidence of the generosity with which the text of

1 See above, I pp. 44-6. 2 Pace GdQ iii, 108.

3 Spitaler, 'Ein Kapitel', 7 no. 39, 9 no. 59. 4 GdQ iii, 69.

5 Jeffery, Materials, 143, 57, 200, respectively; Zamakhshari mentioned both Ubayy and
Ibn Mas'ud, Kashshdf ii, 741 ad loc.

6 Muqatil, Tafsir, MS H. Hiisnu 17, 172'. ' See above, pp. 127-9.
8 See above, loc. cit.; in Ubayy 'sdliha' may be scriptural, Jeffery, Materials, 143.

' Richtungen, 4-32: consonantal, vocalic, additions, synonyms, emendations, 'scribal
errors'; cf. Seeligmann, 'Midraschexegese', 159-60.

204

QURANIC STUDIES

scripture was treated need not be discounted, but the chronological
evolution of the masahif literature strongly suggests that the 'companion'
codices were manufactured from exegetical material in support of an
argument central to the traditional account of canonization, namely, the
'Uthmanic recension. Much of that material persisted anyway in the form
of standard deviations from the canon, accommodated by the ahruf doctrine, 1
evidence in my opinion of a tendency to preserve rather than to neutralize.

With the (later) delimitation of scriptural text and commentary, exe-
getical glosses of the sort represented by 'saliha' achieved a very special
status tantamount to revelation, as may be inferred from Suyuti's observa-
tions on the 'readings' salat al- asrjwa-salat al-'asr ad Q. 2: 238. 2 Citing
Abu 'Ubayd's FaddHl al-Qur'dn, he ascribed the reading without the
conjunctive waw to both Hafsa and 'A'isha, 3 and argued that the function
of the isolated variant (qird'a shddhdha), 4 although basically exegetical,

was of an order higher than that of mere exegesis: SiUJ I ae.\ Jti\ ^ JLyaUl

Jjjtxll Sst^ is jx^ *Jj_/eJI tJ& ^ .L^j L. That the origin of the
reading(s) was not textual but doctrinal, despite the apparatus of trans-
mission, might be inferred from the juridical as well as liturgical signifi-
cance of the 'asr prayer. 5

Now, elevation of the exegetical gloss to the status of 'companion'
reading did not preclude the stigma which might attach to designation of
such as interpolation (mudraj). Critical evaluation of textual variants
{masahif) and/or readings (qird'dt) 6 depended primarily, if not exclusively,
upon attestation (isndd), as may be seen in the relevant nomenclature,
e.g. mutawatir, mashhur, ahdd, shddhdh, etc. 7 Supplementary, but second-
ary, criteria were grammatical feasibility (wajkfil-'arabiyya) and agreement
with the consonantal skeleton (muiodfaqa 'aid rasmjkhatt), the latter appro-
priate in practice only to modification of pointing and of vowels. In
practice, too, the hierarchy of attestation and of transmission (naql) might
include the other criteria, so that the designations mutawatir (generally
attested) and ihiqa (soundly attested) could be in all three respects valid
for the seven major 'readings'. 8 Dam's description of the science of qtrd'a

1 Goldziher, op. cit., 33-51; GdQ iii, 106-8; and see above, I p. 45.
3 Itqan i, 327-8.

3 Apparently correctly, but see GdQ iii, 150 n. 1 ; cf. Spitaler, 'Ein Kapitel', 4-5 nos.
12-20, esp. no. 14, though nos. 12-13 do identify Hafsa with the reading tva-saldt al-
'asr; for the sake of tidiness it may be added that the same reading was ascribed to yet
another wife of the prophet, Umm Salama, see Jeffery, Materials, 235.

* Cf. GdQ iii, 136-7, 155, 228-30.

5 See Goldziher, Richtungen, 14-1 5 ; and id. 'Nachmittagszeit', 294-302.

6 Often undifferentiated, see GdQ iii, 60 n. 2, but cf. Jeffery, Materials, 13-14.

7 See Suyflti, Itqan i, 210-29; GdQ iii, 149-57.

* e.g. Suyuti, Itqan i, 225.

PRINCIPLES OF EXEGESIS

205

^ as one necessarily guaranteed by authority (sunna muttaba'a) may be seen
as theoretical acknowledgement of the primary criterion. 1 Into the con-
ventional lists illustrating the relative ranks of 'readings', 2 examples of
acknowledged interpolation might be fitted by restricting numerically the
extent of attestation: Suyuti's loci probantes included 'min umrri ad Q. 4:
12 and 'ft mawdsim al-hajj' ad Q. 2: 198, and might be thought not very
different from those examples allegedly derived from 'companion' codices
and discussed above. That a considerable amount of exegetical gloss was
not in this way relegated to the masorah, but rather left undisturbed in the
text of scripture, has been proposed. 3

The closed system of 'readings' symbolized by acknowledgement of the
Seven/Ten/Fourteen authorized versions represented a generously defined
consonantal text whose stabilization was chronologically fixed by recourse
to the familiar device of ascription. Illustration of the product of that
method is found in the collection of readings ascribed to Hasan Basri
(d. no/728). 4 The textual standardization exhibited there, e.g. the reduc-
tion of variants to vocalic and diacritic mutation, interspersed with
allegedly 'dialectal' forms, did not suppress altogether the exegetical
moment, as can be seen in yu'bad/tu'bad for na'bud ad Q. 1 : 4, s and kadib
for kadhib ad Q. 12: 18* The role of Hasan as eponym in the elaboration
of a Basran tradition for the science of Qur'an reading was very significant
indeed, despite his presumed insignificance (sic) in that very tradition: 7
it was of course as a source of tafsir that he was selected to be the figure-
head of a regional tradition, a tendency reflected also in ascription to him
of the polemically exegetical Risala fil-qadar. % The polarization round
celebrated figures of originally anonymous dicta, whether in hadith, tafsir,
or qird'dt, reflects an exclusively methodological and tendentiously formu-
lated argument of the Islamic sciences, from which objective historical
data can hardly be elicited.

The same methodological tendency may be detected in the three criteria
(shurut) employed to assess the validity of a scriptural reading: acceptance
for cultic purposes required authoritative and collective attestation (tawa-
turjnaql al-thiqat), which logically presupposed both agreement with
the consonantal skeleton of the 'Uthmanic recension and grammatical

1 Apud Suyutf, Itqdn i, 211; see above, p. 169.
1 e.g. of Ibn Jazari, apud Suyutl, Itqdn i, 215-16.

3 See above, I pp. 27-9.

4 Bergstrasser, 'Koranlesung', esp. 20-46; GdQ iii, 104 n. I, 110.

5 'Koranlesung', 20; cf. above, III p. 108 where the 'reading' was adduced in a quite
different context.

6 'Koranlesung', 36; anonymously adduced by Kalbi, Tafsir, MS Ayasofya 118, I29 r
ad loc.

7 GdQ iii, 165, 177.

8 See above, pp. 160-3 1 the chronological conclusions of GdQ iii, 104 n. 1, are derived
from a historical framework nowhere attested in the collection itself of IJasan's 'readings'.

206

QURANIC STUDIES

feasibility; acceptance for exegetical purposes might waive agreement with
the consonantal skeleton but such precluded cultic employment: Jlsli
Jo Ijii. 1 The distinction was later defined as that obtaining between mash-
hur and shadhdh, terms related, not surprisingly, to degrees of attestation
rather than to functional values. Dispute (ikhtilaf) engendered by contra-
dictory readings or which employed alleged variants as proof-texts, e.g.
yathurna v. yatahharna ad Q. 2: 222, or lamastum v. lamastum ad Q. 4: 43,
could always be resolved by the customary reference to attestation, though
multiple readings were seen by some to reflect the superiority of the
Qur'an over other scriptures, whose revelation (and hence, by a curious
logic, their interpretation) was limited to a single 'aspect' (wajh) : j l^iJl

Option in that kind of dispute was more often expressed by the term
ikhtiyar than by tarjihy-qmte possibly owing to the suggestion of a tertium
comparationis in the latter, which was essentially halakhic. 3 The extent to
which the terms were interchangeable must be qualified in the same way as
employment of the term qiyas by halakhists and masoretes respectively. 4
Acknowledgement of a multiple reading meant in theory recognition of a
multiple revelation (i.e. qiraa was elevated to the status of aya), s an example
of which was 'ajibta v. 'ajibtu ad Q. 37: 12. 6 But an exegete like Zamakh-
shari, preoccupied with textual tidiness, found it expedient to rationalize :
the first-person pronoun must indicate indignation rather than surprise,
or be interpreted as hypothetical, or be annulled by insertion of 'qui
Muhammad'.''

The intrusive character of variae lectiones in the exegesis of Muqatil,
Kalbl, and Sufyan has been noticed. 8 That the Quranic masorah had not
yet been elaborated could be inferred not only from the paucity in hag-
gadic exegesis of textual discussion but also from the elliptical form of
such when found, suggestive of later redactional activity presupposing
masoretic formulations. For example, Muqatil's proposal ad Q. 18: 44
that one may read wilaya (mulk : dominion) or walaya (nusra: support) is
found fully documented in Farra'. 9 Similarly, Muqatil's interpretation ad

1 SuyutI, Itqdn i, 213-14, 225; cf. Ibn Qutayba, Tamil, 32; GdQ iii, 129.
1 SuyutI, Itqdn i, 226-7, cf. also iv, 193-4; *hat argument may be compared with the
antithesis munajjam:jumla wdhida, above, I pp. 36-8.

3 See above, pp. 187-8; GdQ iii, 107, 129-37.

4 See above, pp. 167-9; cf. Suyflti, Itqan i, 229 for tarjih.

5 SuyutI, Itqdn i, 227.

4 See Goldziher, Ricktungen, 21-2: reference T aD an> Tafsir xxiii, 26 ad loc.

7 Kashshaf iv, 37-8 ad loc; the third explanation is an application of taqdir, see below,
pp. 219-21.

8 See above, pp. 127, 132-3, 138.

9 Muqatil, Tafsir, ijo 1 ; Farra', Ma'dni 'l-Qur'dn ii, 145-6.

PRINCIPLES OF EXEGESIS

207

v Q. 18: 80 of khashind (we feared) as 'alimna (we knew) might be thought
to require the documentation of Farra', in which reference was made to the
codex of Ubayy. 1 As has been noted, ascription of variants to 'companion'
codices was characteristic of masoretic exegesis, in which specific authority
was sought for traditions up to that point anonymously preserved and
transmitted. That the equivalence khashind:' alimna for Q. 18: 80 cannot
in fact be derived from the codex of Ubayy is hardly surprising; 2 some
impression of the cavalier treatment of 'companion' codices can be
gained from comparison of the diametrically opposed views of Ibn Hazm
and Ibn Hajar on the exact contents of the codex of 'Abdallah b. Mas'ud
(i.e. the presence/absence of the mu'awuridhatdn). 3

For a textual history of Muslim scripture, as opposed to doctrinal
statements on the formation of the Quranic canon, the parallel passages
which I have described as 'variant traditions' may be thought relevant. 4
Unlike the exclusively exegetical variant readings, glosses, interpolations,
and synonym-equivalences, the variant traditions exhibit at least the com-
ponents of a process by which scripture was produced from revelation.
They represent the only material variants within or outside the canonical
text, and were to some extent so acknowledged in the works of Horovitz
(Koranische Untersuchungeri) and Speyer {Die biblischen Erzahlungen im
Qoran). As might be expected, masoretic employment of this material
was seldom explicit, but one example is Baydawi's reference for the
Zechariah traditions ad Q. 19: 10 (three nights) to the complementary

locution 'three days' in Q. 3 : 41 : ol j±s. Jl J fljVij Ua (JLUI J$"i Uilj

'J^-^S j&UIj jflil ^Ul !»^T & £~J! aJs- ^1 -»J| J* XJ^JOI

^gJUj |»UI- 5 It could, on the other hand, be argued that the masoretes'

use of both textual analogy and periphrasis involved at least implicit re-
cognition of variant traditions, perhaps in the form of similar contexts rather
than as multiple versions of a single narrative. Now, for the transmission
history of Hebrew scripture the role of the Masoretes is often seen to have
been mechanical rather than creative. 6 That they were working within the
(perhaps not so confining) limits of a liturgical tradition cannot, and need
not, be refuted. An essentially consonantal text is, however, susceptible
of a variety of interpretations, semantic as well as grammatical. Establish-
ment of a vocalized text would otherwise hardly have been necessary.
Even in the selection of one of two or more purely orthographic alterna-
tives a degree of understanding, and hence of interpretation, was essential.

1 Tafsir, 172 1 ; Ma'dnl 'l-Qur'dn ii, 157.

2 Cf. Jeffery, Materials, 144; GdQ iii, 88.

3 Suyuti, Itqdn i, 221 ; GdQ ii, 41-2, iii, 179. * See above, I pp. 20-7.

5 Baydawi, Anwar al-tanzil iv, 4.

6 Cf. above, III pp. 100-1 ; Barr, Comparative Philology, 188-222.

208

QURANIC STUDIES

More complex is the character of lingua sacra itself, for which apparently
simple procedures like punctuation and segmentation could, and often
did, involve a doctrinal commitment. For the textual history of Muslim
scripture the activity of the masoretes was not only creative but productive :
of postulates which became the foundations of both grammar and lexico-
graphy. 1 My use of the phrase 'masoretic exegesis' is intended to convey
precisely those creative and productive aspects of the process by means of
which revelation became scripture. The creation of a scriptural apparatus
(masorah) coincided with, or was slightly posterior to, the establishment
of a scriptural basis for Muslim jurisprudence, and represents acknow-
ledgement of the authoritative status of revelation as one source of doc-
trine. The employment of scriptural shawahid in halakhic controversy
required a fixed and unambiguous text of revelation, or at least one in
which ambiguity was conventionally limited. The result was the Quranic
canon.

The major premiss of masoretic analogy was insistence upon the con-
ceptual unity of the Quranic revelation. 2 Formulated as the binary opposi-
tion muhkam.mutaskdbih, that premiss justified comparison of the parts
to one another and eventually of the whole to its constituent parts. In
its initial stages the procedure tended to be self-contained: the kind of
analogy employed was thus deductive. Its basic operations were two:
semantic collation (lexical) and periphrastic restoration (grammatical).
The first of these found rudimentary but eminently practical expression in
a work ascribed to Muqatil b. Sulayman and entitled Kitab al-vmjuh
wal-naza'ir, 3 alternatively K. tafsir wujuh al-Quran and Al-Ashbah wal-
nazir (sic) ft tafsir al-Qur'an* Al-Ashbah wal-naza'ir, 5 Wujuh harf al-
Qur'dn, 6 the last-named adduced as the title of MuqatiTs work from which
the extant recension was made. An undated papyrus fragment of a version
of this work was published and described by Abbott as earlier than the
recension preserved in Beyazit 561, itself (like Topkapi Emanet 2050)
the work of one Abu Nasr, apparently a student of Muqatil but not the
transmitter of his other two exegetical treatises. 7 Neither her reasons for
that conclusion, nor her reference to Muqatil's 'linguistic' tafsir, inspire
confidence. 8 It would be difficult indeed to characterize MuqauTs exe-
getical interests as 'linguistic', 9 and while there may be some connection

1 For the Biblical Masorah Gertner emphasized, in my opinion rightly, the comple-
mentary relation between punctuation/vocalization/accentuation on the one hand, and
interpretation on the other: see 'The Masorah and the Levites', esp. 244-53. His stress
upon the creative work of the Levites may, however, be thought somewhat to restrict
appreciation of the same quality in the activities of the post-Talmudic Masoretes.

1 See above, pp. 166-70, but also 149-53, 155-6-

3 GAS i, 37 no. 3; MS Beyazit 561, Emanet 2050. 4 MS Beyazit 561, i r .

» MS Beyazit 561, 138'. 6 MS Emanet 2050, i r .

7 Abbott, SALPii, 92-7; see above, pp. T44, 172: Hudhayl b. Habib.
• Abbott, SALP ii, 96, 95, 106. • See above, pp. 143-4.

PRINCIPLES OF EXEGESIS

209

, between his teaching and this particular work, I am inclined to regard it
as an essentially independent composition and as having been composed
not earlier than the beginning of the third/ninth century. Appearance
(twice !) of the term ashbdh in Beyazit 561, but not in Topkapi Emanet
2050, poses something of a problem, and may represent an attempt to
classify this treatise with the more or less contemporary mutashabihf
mushtabik lexical collections. 1

The Kitab al-zoujuhwal-nazd'ir (Beyazit 561) contains 186 lemmata in no
recognizable order: conceptual schemata rather than separate lexical
entries. A typical example is the material assembled sub voce wahy: 2

q1 J| L*jf UT jTy3l J** dUI L^jl Cl *}y dUii t LoS/l J* aJUl

cyxlj (6: 19) <u ^Tj^V jTjSJI II* Jl ^jlj Jtfj (4: 163) X$l
J^-J ilj SJuUI J ajji dJUli r L fi J^/l ^ ^jJI JlsJI (2) ^iT

J -UyT' i (5: in) J^-^j ^ jl Odjlj^l ^jl^JI Jl

JL*JI ^ tfj&Jl jf J*JI dbj ^fj Jji. JpdJ! Jl djj ^jtj J*JI

u^Tjj (JT j *jy vbT^yi ajwi a^jIj (3) (16:68)1^
(4) (19: n!) uuj lyw. ji uixT ^Jt <_-^5" J^5j p^Ji
cU^, js* (j (^jij s^»-Ji ti ^1 ^^ii ^-'^ *^L?

JI ^Jlj ^1 jjJ^Ui f U»"Sfl J Jlij (41: 12)

(^JaL^JI jl_J j»Ui it ijj*, j Jl»J (6: 112) L^jju ^aj o j^itj J^5j (jOJO

Af^bCs) ( 6: I21 ) u^j^'j '^J-J^ J' o>»^5

w <j»J otj j>/i\ ^-JjJj i^i j *Jy «uu Jyji <^>ji (^wi

(99 : 5) W (Jls- Thus, for five allegedly distinct uses of the term wahy,
nine scriptural passages were adduced:

1. Wahy: revelation (Q. 4: 163, 6: 19)

2. Wahy: inspiration (iUtam) (Q. 5: m, 16: 68)

3. Wahy: writing (kitab) (Q. 19: 11)

4. Wahy: command (arm) (Q. 41: 12, 6: 112, 6:121)

5. Wahy: speech (qawl)(Q. 99: 5)

Selection of the verses was arbitrary: of seventy-two occurrences in the
Qur'an of a finite form of the verb atvha and six of the substantive wahy,
not more than five, possibly six, separate contexts can be elicited, and those
correspond roughly to the five 'aspects' (wujuh) of Beyazit 561. But some
qualification, as well as differentiation, is necessary. The first aspect

1 See below, pp. 212-16. 2 MS Beyazit 561, 94 r -s v .

2IO

QURANIC STUDIES

(revelation) is, for example, logically misleading since, as a descriptive
term for divine communication, it must include the remaining four
aspects. Strictly interpreted, the definition of revelation as 'that mediated
by Gabriel to the prophets' cannot be reconciled with Muslim doctrine
on the modes of revelation, but if that difficulty is overlooked, wahyjawhd
as 'revelation' will account for a quarter of its Quranic occurrences (of
which seven refer to Moses, and five to Noah). 1 For the second aspect
(inspiration) the traditional locus probans was indeed Q. 16: 68; inclusion
here of Q. 5: 111 presupposed a sophisticated distinction between pro-
phets and apostles, but one well established in the tradition of Judaeo-
Christian polemic, possibly the source of the semantic equivalence wahy:
ilham. 2 Logically, the passages alluding to demonic inspiration (Q. 6:
112, 6: 121) belong here and not under the fourth aspect (command/
decree), which exhibits wahy as manifestation of the divine will in creation.
For that function of wahy, Q. 41: 12 is appropriate, as would have been
inclusion of Q. 99: 5. Use of the latter to illustrate the fifth aspect (speech)
is curious. In fact, a separate function of wahy as speech is only signifi-
cant as one component of the contrast between the fifth aspect and the
third (writing), for which the choice of Q. 19: 11 might seem singularly
inappropriate. 3 Incidentally, the confusion here between Suras 3 and 19
seems to reveal uncertainty about the place of variant traditions in the
canon. 4 That for the third aspect kitab cannot be 'decree' seems clear
from the separate listing of the fourth aspect. 5 The intended reference
may have been to wahy as scripture, for which Q. 17: 39, 18: 27, 20: 114,
29: 45, etc. could be proposed. And Q. 19: 11 would illustrate wahy as
speech. Absent from the lexical analysis in Beyazit 561 is the notion of
wahy as dispatch (irsdt), for which Q. 42 : 52 might have been adduced. 6
The polysemy of wahy in the document of revelation is thus attested, but
in a manner clearly anterior to the elaboration of sophisticated semantic
analysis.

That example is characteristic of the 186 lemmata: with the exception
of eleven items (the particles: li, siwd, hal, ft, win, aw, am, fawqa, ma,
hatta, and ilia) all may be described, within the framework of lingua
sacra, as theological concepts, e.g. hudan, kufr, din, ithm, dalal, sirat, etc.
But the principle by means of which the several uses of each were differen-
tiated was of broader linguistic application: namely, the reference to con-
text. However inappropriate the actual loci probantes might seem, the
number of aspects (wujuh) for any particular item approximates very closely
to the number of scriptural contexts in which the locution appears. The

1 See above, I pp. 33-8. 2 See above, II pp. 58-9.

3 Thus Ibn Qutayba, Ta'zvil, 373-4 s.v. wahy.

4 Cf. e.g. Baydawl on the same subject, above p. 207.

5 See above, II pp. 76-7.

6 That sense was adduced by Ibn Qutayba, Ta'toil, 373-4, using Q. 6: 19.

PRINCIPLES OF EXEGESIS

211

lexical data thus isolated might more accurately be described as reference
rather than as information. 1 This kind of lexicography was exclusively
concerned to elucidate scriptural imagery, and hence avowedly exegetical.
The type, exemplified in the primary title of Beyazit 561, K. al-wujuh
wal-nazd'ir, became a genre of exegetical literature and a basic com-
ponent of the masorah. Ascription of its origins to Muqatil was occasionally
explicit. 2 MalatT, too, in his K. Tanbih ascribed to Muqatil a series of
semantic equivalences in which much of the scriptural lexicon was reduced
to a limited number of standard synonyms. 3 But the method employed
there was the obverse of Muqatil 's 'aspects': only one 'meaning' was ad-
duced for each scriptural concept in whatever context (cf. the introductory
formula: hull shay' fi'l-Qur'dn . . .ya'ni. . .). The material includes most of
what I have designated minimal units of explication, e.g. kadhdlika : haka-
dhd, ladayna.'indam, la'alla:likay* but also several which presupposed a
rather more complex interpretation, e.g. khdtam.tab', ghuluf.akinna. 5
In that arrangement the__informing principle was not 'aspect' (wajhj
wujuh) but 'analogue' (nazirjnaza'ir). Both methods are illustrated in
Suyuti's chapter on al-wujuh wal-nazair, in which the nazd'ir (kull shay' ft
'l-Qur'dn) were qualified by specifying the exceptions to their general
'meanings'. 6 Thus gradually, a more or less fixed pattern of semantic
distribution was elaborated for the lexicon of scripture which included
both polysemes and/or homonyms (wujuh) and synonyms (nazd'ir). 7
Though the early stages of the process exhibit a concern primarily with
substantives, the dozen or so purely grammatical elements in Beyazit
561 generated a similar but separate treatment of particles (addtjadawdt),
which in Suyuti's treatise ranged from the interrogative hamza to the
vocative yd.*

The earliest uses of the term nazir vary: 9 in the works ascribed to
Muqatil the 'analogy' is explicitly textual, e.g. for bdkhi'un nafsaka (Q. 36:
3 ad 18: 6), for shatatan (Q. 38: 22 and 72: 4 ad 18: 14), and for saffan
(Q. 20: 64 ad 18: 48). 10 In Bayazit 561 the term nazir occurs passim,
always, and not unexpectedly in view of the nature of the collection, to
introduce an analogy based upon an identical word or phrase. In both
H. Hiisnii 17 and Beyazit 561 the term was employed interchangeably

* See above, p. 142, and III, pp. 99-100. 2 e.g. Suyuti, Itqan ii, 121.

3 See above, pp. 165-6; Malawi, Kitab al-tanbih, 56-63.

4 See above, pp. 129-30. 5 Cf. above, II pp. 64-5 and 72-3, respectively.

6 Itqan, now' 39: ii, 121-39(122-31 and 132-9, respectively); these specifically masoretic
terms are not to be confused with the later juridical ashbdh wa-nazd'ir, though the prin-
ciple of analogy underlying them is of course the same, cf. Schacht, Introduction, 1 14, 265.

7 The type had an ancient pedigree from the Hellenistic schools of rhetoric and became
eventually a source for both halakhic and masoretic exegesis, cf. Daube, 'Rabbinic
methods', 241 n. 7; and Wurthwein, Text, 21-2 (e.g. Okhla we-okUa).

8 Itqan, naiv' 40: ii, 140-239- 9 above, pp. 127, 143, 169-70.
10 Muqatil, Tafsir, MS H. Hiisnii 17, i67 T , i68 r , i7o c , respectively.

212

QURANIC STUDIES

with tnithluha and ka-qawlihi: 'textual analogies' were derived exclusively
from the text of scripture. The significance of that method emerges from
comparison of Muqatil with his contemporary Sufyan Thawrl. The latter's
use of analogy, introduced not by nazir, but rather by ka-qawlihi or
mithlu (qawlihi), was limited to scriptural shawahid but not to litteratim
constructions. For example, the connection between Q. 2: 28 and 40: 11
derives from the fact/assertion of resurrection but not from linguistic
expression of that fact/assertion in those passages. 1 Similarly, acknow-
ledgement of divine creation is the only link between Q. 3 : 83 and 43 : 87,
and no attempt was made to elucidate the grammatical and lexical charac-
teristics of one passage by reference to the other. 2 In the work of Sufyan,
as it has been preserved, masoretic material is unmistakably intrusive, and
the same may be said at least of Muqatil's Tafsir (H. Hiisnii 17). 3 The
relationship of Beyazit 561 to Muqatil is, as has been noted, problematic:
its contents, like the lexical data ascribed to Muqatil by Malati, signal the
beginning, rather than the end or a stage along the way, of the exegetical
development which I have called masoretic. With that development the
term nazir came to designate syntactical/grammatical analogue generally,

and not merely with reference to the language of scripture. For example,
in Farra' Q. 49 : 1 1 ) | of ^j**- was proposed as nazir to the
construction with an auxiliary verb in 18: 31 t_,|^ij| The analogy

was thus based upon scriptural usage but not upon an identical phrase.
Similarly, Zamakhshari introduced with nazir an analogy to Q. 4: 48 <UJ I j]
e.UL> ^yJ iiMi ijji L jJoj Aj ^j^i V from the wus loquendi

(qawluka) t Uj ^ jlk&)| Jjuj jUjJOI Jjuj V j~>Vl oj by means of
which the ellipsis in the scriptural passage could be resolved. 5 In Farra'
are found also the locutions mithla, ka-qawlihi, and wa-hiya bi-manzilat
qawlihi, but more often than not to introduce analogies based upon identical
wording. 6 It was that method which remained characteristically masoretic
and in which exegetical use of the term nazir had originated.

From delimitation of homonyms/polysemes and synonyms it was but
a short step to a distributional analysis of Quranic diction. Its formative
principle was designated mushtabih (variant: mutashabih)? The earliest
collection of data organized according to that principle is ascribed to the
masorete and (seventh) 'canonical reader' Kisa'i (d. 189/804). 8 The work

» Sufyan, Tafsir, 3. * Sufyan, Tafsir, 37.

3 See above, pp. 132-3, 138. 4 Ma'dni 'l-Qur'an ii, 141-2.

5 Kashshdf i, 519-20: an alternative to his resolution of the construction by taqdir,
cf. Wansbrough, 'Periphrastic exegesis', 259, and below, pp. 219-24.

6 Ma'dni 'l-Qur'dn ii, 137, 157, 5, respectively. 7 See above, pp. 157, 165.
8 GAS i, 17, 48; MS Beyazit 436, entitled Kitdb mushtabihdt al-Qur'dn, but given 70'

as Kitdb al-Mutashdbih; see also GdQ iii, 180, 188, and passim.

PRINCIPLES OF EXEGESIS

213

\ consists of serial enumeration by sura of locutions which occur once,
twice, and from three to ten, 1 fifteen, and twenty times in the text of the
canon. Appropriate illustration is provided by the material assembled
under the heading 'once' for Surat al-Baqara: 2 • S^SJl Sjj^. ^ J^LJI i_jU
(j* ajj^i \jli\3 [$Jj j^SCj Ijjul jTjSJI jjLj(2: 21) j»5^j IjJ^aI ^UJI L^jL
J Lj (2: 23) j^TtU^-i Ifeij & Jt M 4i* jTpUl jj'Lj (2: 23) Ail,

jT^JI y'Lj (2: 41) t UI iny^ cJ_yl Lj UuIj l&Jj j^JzxJ ^ (jj^Lil

^jjlj l^Jj Jp. ^ij Q\ J i}\ JLj (2: 6l) (JPtll jjo l$Jj i)_^a^»

L^J t£jl^JI JJ O^iUaJlj jT/31 ^Lj (2: 62) ^UaJlj (^jlyjJIj |_J.iU

V L>J> objJu^ oT^I JL»j (2: 80) Jojju., UUf VI jUI Lu^J
Lvij OjJ^i Vj jT/31 JL>j (2: 86) jj^j j^a (»4^
^ t^oJI jjo I4JJ 4JU jTyJI J (2: 100) o>->Jj V ^ji^f

dJUj^U UIuj U>ij lie.U U Jjo oT/JI ^*Lj (2: 120) pLJI
pa«^«!i p&fjiJ oTjiJ! yLj (2: 129) pfcjT^jj iJC>t)lj ojLOI
JaI Lj LjJj cb jsij jT/JI jjLj (2: 150) e.Lj Jij^U-lj ^j^sci" }Ai
■uL ^'1 jL Vj L>ij 4j -UJ( ^JL) jTjil! yLj (2: 173) <U)| jjj
^ (2: 185) jSi Lu^ jfc' I4J5 <dt. oLill J ^Jjfc: 173)
jT/Jt jjLj (2: 218) Ij^L ^jJlj ly-T ^jdl jl L>i> «di-> oT^JI J
jT/JI yLj (2: 271) ^SoL;*. ^ l&ij ^-iJb *M b^-^j

oL*)l jjLj (2: 170) Uc-U! jJi* LJJI L LgJj ^yi jJo ^5d'

b («'c) cisj l$Jj !)*^3 (2: 35) iJij L^u L>ij Uj^j

. &u oT^SJ! J (2: 233) ^ lJ1£j V L>Jj (2: 35) dLxjjj cJf

In this passage twenty-one instances of 'unique' phraseology were
adduced, and in all but one (the second example for Q. 2: 35) their unique-
ness either reiterated (laysafi 'l-Qur'dn mithlukujghayruhu: 2: 100, 2: 173,
2: 185, 2: 233) or contrasted with what is apparently 'normal' Quranic
usage (wa-sd'ir al-Qur'an . . .). The contrasts thus established are not
semantic but grammatical, and turn upon the presence/absence of particles
and prepositional phrases, variation in word order, in inflexion, and in
orthography. As in the lexical collations of Beyazit 561 and similar works,
the shawdhid are entirely scriptural, and no effort was made to justify a
particular construction by reference to usage outside the Quran. On the
other hand, the selection of contrasts might be thought arbitrary: the
locution u'budu rabbakum (Q. 2: 21) appears also in Q. 22: 77 and, in a
slightly expanded form, in 5 : 72 and 5:117. If it is the apostrophe yd
ayyuha 'l-nds which is here operative, that appears in combination with

1 Read so, for 'eleven', 6i T .

1 Kitab mushtabihat al-Qur'an, i T -3 r .

214

QURANIC STUDIES

ittaqu rabbakum only three times (Q. 4: 1, 22: 1, 31: 33), from which a
stylistic norm can hardly be elicited. Rather more valuable would have
been reference to the contrast between Q. 2: 21 and the collocation of
yd ayyuha 'lladhina amanu as the mode of address with the imperative
ittaqu 'lldh(e.g. Q. 2: 278, 5: 35, 9: 119, 33: 70, 49: 1). But isolation and
comparison of formulaic phraseology as an exercise in literary analysis
was not the author's purpose, though the material assembled here may be
useful to that end. Acknowledgement of variants in the form of inflexion
(Q. 2: 62, but also 22: 17; only 5: 69 has casus rectus) and of orthography
(Q. 2:150, but the alternative spelling occurs only twice, in 5 : 3 and 5 : 44)
exhibits concern for the conceptual unity of scripture, noticed above for
the lexical category of wujuh. That acknowledgement emerges very clearly

from the inclusion here of Q. 2: 233, whose locution (JibsJ j and

attendant pointing with the internal passive may profitably be compared

with the occurrence of LiS Lj&So/iJiiSi V in 2: 286, 6: 152, 7: 42, 23:

62, 65: 7, but also 4: 84 duiJ VI tJlSC> V- Like the variant traditions,

of which these instances represent the formal aspect, the textual variations
preserved in the canon might be thought to throw some light on the forma-
tion of Muslim scripture. 1 In KisaTs work the principle of mushtabihat
was sound, its application fragmentary. Subsequent elaboration of the
genre provided more complete coverage of the phenomenon. 2 It may
plausibly be argued that the collation of mushtabihat reflects awareness, on
the part of the masoretes, not only of the Qur'an's stylistic homogeneity
but also of its structural idiosyncrasies, and finally, of the necessity to
explain these in terms of intrinsic analogies.

An example is Q. 2: 35 4^>JI tiUjjj oif t j5Cj joT L Ulsj

j^^JLkJI ^ Uj£ji 'ij>z£J\ aO* LjSi' Lxi*i incompletely and

incorrectly adduced by Kisa'i. That the verse is a variant of the Adam
tradition in Q. 7: 19 ^ aj^JI tiLijjj cj! t jSCj bj

^JuaJI ^ UjJci o^9t£JI Lyu "jj U^Li hardly requires demonstra-
tion. Juxtaposition of the two verses, not as masahif but as mushtabihat,
generated the following exegesis: because in the first verse the divine
command was to 'dwell and feast', the co-ordinating conjunction (wa-)
was employed to link the two actions as one divine favour, the limitless
extent of which is emphasized in the terms 'copiously' (ragkadan) and
'wherever you like' (haythu sWtuma) ; in the second verse the command
was to 'take (up) residence, then nourishment', separate actions requiring

1 See above, pp. 207-8.

2 See Suyuti, Itqdn, nana' 63 : iii, 339-44, and the late works mentioned there, 339,
to which may be added Raghib Isfahan!, Hall mutashabihdt al-Qur'dn.

PRINCIPLES OF EXEGESIS

^ to be temporally distinguished by the subordinating conjunction {fa-),
its restricted largess expressed by omission of 'copiously' and by
insertion of a limiting preposition {rnin) before 'wherever you like'. 1
Whatever suspicions that interpretation might provoke, it ought to
be quite clear that the presence of doublets in the text of scripture
could not be an embarrassment. But all such material, absence of
which could reduce the size of the canon by forty to fifty per cent, lends
itself admirably to the kind of documentary analysis proposed for the
Shu'ayb traditions and the 'double-garden' imagery. 2 That in the Muslim
tradition recourse was had not to documentary analysis, but to exegesis,
may be illustrated by considering the number of Mu'tazilites responsible
for works in the genre of mushtabihdtjmutashabihdt. 1 An element of rhetori-
cal criticism is also evident, at least in the later development of the genre.
Repetition in scripture came to be described and evaluated in specifically
aesthetic terms, and textual variation regarded as stylistic embellishment.
The technical vocabulary formulated to that end, e.g. munasabajirtibdt
(filiation) and tafannun (elegant variation), presupposed the structural
integrity of a single document of revelation. 4 That terminological refine-
ment, not yet expressed in the work of Kisa'I, reflected the doctrine of
i'jaz al-Qur'an, a post-masoretic phenomenon. Once acknowledged as an
appropriate object of rhetorical analysis, the text of scripture was safely
removed from the danger of dissolving into its original and fragmentary
components. 5

The exegetical procedures symbolized by the terms zoujuh, nazair, and
mushtabihdtjmutashdbihdt were derived from a view of scripture as self-
contained and self-explanatory. The logic of that view rested implicitly
upon acknowledgement of lingua sacra as a special mode of communica-
tion. The schemata of revelation were, so to speak, sui generis and could
gain little or nothing by reference to the elements of normal linguistic
usage. To describe that mode of communication as the 'word of God'
might be thought perversely dogmatic, but points none the less to the
fundamental distinctiveness of the literary expression which is the subject
of these studies. The manner in which originally or basically neutral
elements in language (if such can ever be said to exist) achieve separate
reality as the ingredients of fixed and traditional imagery (whether or not
'scriptural') is well known: all 'meaning' is related to context. Thus,
the most recent attempt to analyse the lexicon of Muslim scripture derives
from a semasiology not appreciably different in kind from that underlying
the 'aspects' of Beyazit 561: namely, Allard's Analyse conceptuelle du

1 Suyuti, Itqdn iii, 340. 2 See above, I pp. 20-7.

3 See GAS i, 618-19, 62a, 636, but also 13, 44.

* Suyuti, Itqdn iii, 340, 342, and nato' 62: iii, 322-38.

5 See below, pp. 227—46, and above, II pp. 77-83.

2l6

QURANIC STUDIES

Coran sur cartes perforees (1963).' Though the author was unable to resist
including in his 430 cards some irrelevant information (e.g. the 'chrono-
logy' of revelation under (B) Cadre), the results which can be obtained
from manipulation of his system are comparable to those elicited from a
literary analysis (recurrent phraseology, as opposed to separate lexical
items, is unfortunately accessible only through non-literary headings:
anthropologie, thiologie, Mdque-religion). To establish, quickly and con-
veniently, statistics on conceptual distribution, the method is admirably
suited; to anyone at all familiar with Muslim scripture the results are
invariably predictable (e.g. quantitative emphasis upon the 'Mosaic
syndrome' in Quranic prophetology). 2

Now, in the Muslim exegetical tradition efforts to clarify the lexicon and
imagery of scripture were not always confined to the material of the
document itself. It is difficult, if not altogether impossible, to determine
chronologically the point at which the problems of lingua sacra could be
fruitfully referred to the data of profane literature. From what appear to
be its earliest attested stages, the procedure was at first essentially lexical
and methodologically atomistic. Lexical treatment in haggadic exegesis
seldom consisted of more than a straightforward equivalence adduced
without authorities, occasionally of a foreign etymology for exotica and
hapax legomena, both practices hardly altered by the halakhists. 3 For neither
may sporadic reference to 'Abdallah b. 'Abbas be understood as appeal
to an authority especially qualified in the sphere of lexicology. And yet,
the origins of literature concerned specifically with the scriptural lexicon,
not solely as the expression of theological concepts but as communication
drawing upon the resources of a national language, are almost always
connected with his name. 4 At least three tides of such works have been
preserved: Kitab ghartb al-Qur'dn, KitabjBayan lughat al-Qurdn, and
Masa'il Nafi' b. Azraq. s The substance of the material designated by these
titles has been transmitted in several scarcely varying recensions, and is
synoptically accessible in Suyuti's Itqdn. 6 Whether or not the Berlin MS
Petermann II, 405 is an extract from Suyuti, may, in view of the wide-
spread transmission of the material and of the legendary stature of Ibn
'Abbas, be thought quite without significance. 7

The collection of lexical explanations known as Masa'il Nafi.' b. Azraq
exhibits an exegetical method considerably posterior to the activity of

1 Cf. also Allard, 'Une Methode nouvelle', 5-21 ; a primitive, because dependent upon
non-linguistic data, application of 'contextual semantics' may be seen in Izutsu, Ethico-
religious Concepts.

2 Allard, 'Une Methode nouvelle', 19.

3 See above, pp. 124, 143, and 181, 182, respectively.

* See Goldziher, Richtungen, 69-71. * GAS i, 27-8 nos. 2, 4, 3, respectively.

6 Itqdn, nam' 36: ii, 6-46, 47-54, 55-88.

7 MS Petermann II, 405, 93-101, and Itqdn ii, 55-88; cf. GAS i, 27 no. 1 vs. GAL,
Suppl. I, 331, and Mittwoch, 'Ahlwardt No. 683', 339-44.

PRINCIPLES OF EXEGESIS

^ Ibn 'Abbas (d. 68/687) : namely, the reference of rare or unknown words
in scripture to the great corpus of early Arabic poetry. 1 That method was
in fact so conscientiously and consistently applied in the Masd'il as to
provoke the question whether the real purpose of the work was not to
furnish an ancient and honourable pedigree for what became, with the
masoretes, a very important exegetical principle. Lexicology, like the other
Islamic sciences, was ultimately denned in terms of traditional authority,
with the customary reference to the linguistic competence of the com-
panions of the prophet : what they did not know could not be known. 2
That one of those companions, Ibn 'Abbas, should be able for each of 190
Quranic locutions 3 to cite a verse from Jahili or Mukhadrami poets (many
anonymous) was indeed an accomplishment worthy of rote. Suspicion of
a tour de force is corroborated by appearance of the same lexica in other
scriptural vocabulary lists, also ascribed to Ibn 'Abbas but without
poetic shazodkid.* That principle of which the Masd'il represent an almost
polemical expression was explicitly articulated in dicta attributed to Ibn

'Abbas : d\ji * j*-^ I (Poetry is the register of the Arabs, scil. of their
language) 5 or, in more detail, to 'Umar b. Khattab: SJaUJI jUuiL j^CJL*
lxT"j>,..iI I4J <jU (Learn the poetry of the Jahiliyya, for there you will

find the interpretation of your scripture). 6 Similar exhortation, but here
chronologically unexceptionable, was ascribed to Tabari, Sharif Murtada,
and Jubba'i, 7 by whose dates the practice was well established. That it was
not so prior to the third/ninth century is, in my opinion, very significant. 8
A virtual terminus a quo may be elicited from Ibn Hisham's recension of
the Sira : e.g. for bakhi'un nafsaka in Q. 18 : 6 a line from Dhu Rumma was
adduced, for shatatan in 18: 14 a verse from A'sha. 9 Application there was
exclusively lexical, and thus provides a neat contrast to the haggadic
method of dealing with the vocabulary of scripture. In Bukhari's K.
Tafsir only one line of poetry (anonymous) was cited, for la-awwdkun
in Q. 9: U4. I0 In Muslim a single verse was adduced, at Q. 7: 31, 11 and in
Tirmidhi none in an exegetical sense. 12 Poetry was very occasionally cited,
for lexical explanation, in the works of the halakhists, such as Jassas and

1 See above, pp. 142-3, and III, 97-102. 2 e.g. Suyuti, Itqdn ii, 3-4.

3 Not 140, as in Mittwoch, op. cit. 342.

4 e.g. Suyuti, Itqan ii, 6-54, and see below, pp. 218-19.

s Suyuti, Itqdn ii, 55; cf. Goldziher, Richtungen, 70 esp. n. 3: there is on the contrary
every reason for not accepting the authenticity of that report.

6 Goldziher, op. cit. 69 n. 4. 7 Goldziher, op. cit. 92, 116, 130, respectively.

8 Pace Noldeke, BSS, n n. 6; cf. GdQ ii, 192 (revised and appropriately sceptical).
» Ibn Hisham, Sira i, 302, 304. 10 Bukhari, Sahth iii, 248.

1 1 Muslim, Sahth viii, 244.

12 A line ascribed to the prophet was incidentally included ad Q. S3 : 32, in Tirmidhi,
Sahth xii, 173, a gratuitous insertion comparable to the line from Hassan b. Thabit ad
Q. 24: 15, in Bukhari, Sakih iii, 297.

2l8

QURANIC STUDIES

Ibn 'Arab!. The earliest exegetical composition in which poetic shawahid
were regularly employed is the Ma'dni 'l-Qur'an of Farra' (d. 207/822). 1
But there application of the principle was not limited to lexis: grammatical
phenomena were also justified by reference to profane literature. For
example, the locution uoa-iqam al-salat in Q. 24: 37 required, in the view
of Farra', to be explained as an apocopate permitted by status constructus,
i.e. for iqamat al-salat. An anonymous verse containing 'ida 'l-amr for
'idata 'l-amr was adduced in support of the contention. 2 The technique
could be extended also to syntactical problems, accessory to the peri-
phrastic principle known as taqdir/majdz. 3

In the related field of profane lexicography a parallel and contem-
poraneous practice has been noted: in Khalfl's Kitab al-Ayn verses of
ninety-nine poets together with some anonymous lines were adduced in
support of the usus hquendi* For the scriptural lexicon, however, there
appears to have been some opposition to that method. 5 One title recorded
for the collection otherwise known as Masail Ndfi' b. Azraq is Kitab
gharib al-Qur'an, 6 but a separate work, also entitled K. gharib al-Qur'an
and also ascribed to Ibn 'Abbas, does not contain the Masd'il and does not
employ poetic shawahid. 7 It is instead a plaidoyer for the exclusively Arabic

vocabulary of scripture, an argument set out in the following preface: 8 ^s.

^ tjjs- J*- & jij u^J jUJj JlS jUL <U)I J (j-Lt £j|

t^Sslj Ljjj vyJ' **J *s» lAja"j Vj*" o^J ^ ^.j!

^ jJiJUej V ^j*9 ^jJ^tJIj Jwl UU oUUl. The conflicting interpreta-
tions of Q. 14: 4 and the ambiguity of the locution lisdn 'arabi have been
remarked. 9 A sentiment similar to that of Ibn 'Abbas was attributed to
Sufyan Thawri: -s^U^J* ^ iu^U VI ^yj ^i,™ while

Muqatil reported, on the authority of Sa'id b. Jubayr that all languages
were represented in the Qur'an: jl/Ji <j <UII l$J>)'t Vl 2UJ J 1— 11

The latter view was amplified by Suyiiti to assert the universality of Mu-
hammad's mission: jlj J* ^j(Q. 14: 4) . . . ij JS" J| ^| i^f,

1 See above, p. 150.

2 Ma'am 'l-Qur'an ii, 254 ad loc.; cf. Vollers, Volkssprache, 156-7.

3 See below, pp. 223-4. 4 Wild, Das Kitab al-'Ain, esp. 42-51.

5 Cf. Suyiiti, Itqan ii, 55; and see below, pp. 223-4.

6 Mittwoch, 'Ahlwardt No. 683", 341.

» GAS i, 27 no. 2; MS Atif Efendi 2815. « MS Atif Efendi 2815, io2 r .

9 See above, II pp. 53, 81, HI pp. 98-9.

10 Apud Suyuti, Itqan i, 130.

" Muqatil, Tafsir, MS H. HUsnii 1% 17? at the end of Surat al-Kahf; cf. Itqan ii, 106.

S

PRINCIPLES OF EXEGESIS , 319

\ yt, 'iiL aI^J (jg jlj j»y JS* O^J ^ 4j dj^JI cj^' J OjSo. 1
Some portion of the contradiction was neutralized by resort to the notions
of coincidence between languages {tawdfuqjtawdrud al-lughdt) and assimi-
lation (ta'rib) ascribed to Ibn 'Abbas and Abu 'Ubayd. 2 The concept
of 'pure Arabic' ('arabi mahd) forcefully asserted by Abu 'Ubayda became
an axiom of masoretic exegesis. 3 The link with Ibn 'Abbas was, however,
maintained. The third work of lexical character ascribed to him is entitled
Baydn lughdt al-Qur'anjal-Lughat fi 'l-Qur y dn,* at least one manuscript
version of which contains the same material as K. ghartb al-Qur'dn, with
an identical preface. 5 They are a compilation by sura of standard lexical
explanations, unaccompanied by authorities or loci probantes, either from
scripture itself or from profane literature, and similar to the list of such
transmitted from Ibn 'Abbas via Tabari by Suyuti. 6 The central position
occupied by Ibn 'Abbas in the lughdtjgharib literature emerges from exami-
nation of the data collected by Sezgin for those titles: we are confronted
not by several independent traditions but by scarcely discernible variations
of a single tradition. 7 The lexical data extracted from Bukhari and alpha-
betically assembled by 'Abd al-Baqi contains the same standard material,
and belongs to the same collective tradition. 8 It was not until elaboration
of the genre by Ibn Qutayba (d. 276/889) and Sijistani (d. 330/942) that the
traditional stock was refurbished, substantially and methodologically.'

More important for the Quranic masorah than either variant reading
or lexical explanation was the analysis of grammar and syntax represented
by the exegetical principle taqdirjmajdz. The earliest formulation of that
principle is found in the work of Abu 'Ubayda (d. 209/824) entitled Majdz
al-Qur'dn. 10 In an introductory chapter the author enumerated thirty-
nine kinds of majdz occurring in the text of scripture, illustrated by sixty
shawdhid, ten of them adduced twice, to different ends. 11 Abii 'Ubayda's
typology includes six categories of grammatical and syntactical phenomena,
of which three contain solutions proposed to more or less straightforward
textual problems, i.e. lexis (no. 28), varia lectio (nos. 27, 29, 38), and con-
cord (nos. 4-13). Characteristic of the last-named category are explanations

1 Itqdn ii, 107.

2 Cf. Suyuti, Itqdn ii, 105, 108-19; see Kopf, 'Foreign words', 191-305, esp. 202-4;
and id., 'Religious influences', esp. 34-8, 40-5.

3 e.g. Abu 'Ubayda, Majdz al-Qur'dn i, 8, 17.

4 GAS i, 28 no. 4; MS Esad Efendi 91; I have not seen Munajjid's edition of the
?ahiriyya MS (Cairo, 1946).

5 i.e. MS Atif Efendi 2815; this preface is at i04 r of MS Esad Efendi 91.

6 Itqdn ii, 6-46. 7 See GAS i, Indices: Buchertitel.
8 'Abd al-Baqi, Mu'jam ghartb al-Qur'dn, 1-233.

• GAS i, 48 and 43-4, respectively.

10 GAS i, 48 ; ed. Sezgin, see Wansbrough, 'Periphrastic exegesis', 248.

11 Abu 'Ubayda, Majdz al-Qur'dn i, 8-16: set out with translation in Wansbrough, op.
cit. 248-54 (where the locution avila lakafa-awld, Q. 75 : 34, under nos. 17 and zo, may
better be rendered 'Woe unto thee, woe', see Zamakhshari, Kaskshafiv, 664 ad loc.).

220

QURANIC STUDIES

involving a constructio ad sensum, employed to neutralize contradiction
between formal and conceptual reference to number, person, and gender,

e.g. Q. 22: 5 (no. 4) (^UW) ^iL In another category, ellipsis

(nos. 1-3), the same concept was employed to justify periphrastic restora-
tion. One example of the latter deserves notice, Q. 12: 82(no.2)( ( JjtJ) J.»j
*1b» UUf ^Jl (J j I4J liT^JI 5jj«JI. Insertion of 'the inhabitants
of before 'village', and of 'those in' before 'caravan' is not really essential
to understanding the passage, but indicates, rather, the author's conscious-
ness of a metaphorical construction. It is thus quite unlike the other
illustrations of ellipsis included there (Q. 38: 6, 2: 26, 39: 73) and quali-
tatively distinguished from the examples of irregular concord. But a further
category of majaz (nos. 14-23, 26) contains a number of rhetorical pheno-
mena, e.g. fictio personae, apostrophe, chiasmus, from which together
with Q. 12: 82 it would be tempting to infer that for Abu 'Ubayda the
term majaz designated figurative usage. That it did not, however, seems
clear from the bulk of his 60 loci probantes, and in particular for the
category of idiom and/or solecism exhibited in nos. 24, 25, 30-7, most
especially no. 39. Here majaz represents the rationalization of careless
style, of syntactical ambiguity, even of grammatical error, by means of
restoration according to the norms of scriptural usage.

The method is, at least implicitly, that underlying the type of analysis
identified by mushtabthdt. 1 The functional confusion between nomen
regens and nomen rectum exhibited in Q. 28: 76 (no. 24) and 2: 171 (no. 25)
required justification: occurrence of the second example in the text of
scripture is fairly frequent. 2 Variable function of particles (no. 30), as in
Q. 2 : 26, 20: 71, 83 : 2, and 43 : 51-2, was explained by a series of arbitrary
but not quite irresponsible equations, e.g. fatoqaha.dunaha, ft:' aid, etc.

On the other hand, Q. 79: 30 UU: diii (^*) Juo o^J^U involved a
dogmatic postulate relevant to the chronology of divine creation and was
recognized as one of the 'standard puzzles' of scripture: 3 substitution of
'together with' for 'after' was thus, in the precise sense of the term, exe-
getical. The presence of particles (nos. 31-3) was in the examples here
adduced (Q. 83 : 1-3, 1: 5, 16: 98) exclusively grammatical, and in the
first two majaz represented the resolution of synthetic constructions. With
regard to the third, it might well be argued that the absence/presence of bi
after qarda, contrasted in Q. 16: 98 and 96: 1, is not merely not optional
but reflects, indeed, a semantic distinction. Conversely, the majaz proposed

for Q. 20: 69 (no. 36) ^>.L xS \^*^> jl) \yi^> LSI indicated a genuine,

and frequent, option, which may be regarded as stylistic rather than

1 See above, pp. 313-15. * See above. III pp. 1 13-41.

3 As, for example, in Muqatil, Tafsir, MS BM Or. 6333 ; see above, p. 164, no. 3.

PRINCIPLES OF EXEGESIS

221

grammatical. Common, or optional, gender for collectives (no. 34) can be
demonstrated by reference to Q. 26: 105 ^LyJI ^ji coif
but not of course to 16: 66 Aijiu J 1^, ^.^.J o^jj j.UiSl J ^ j]j
(LgJjiu). Finally, the incondnnity exhibited in Q. 73: 18 (no. 35) C L*JI
a> JaJc^ (uiuJI) and 55: 19 ... 22 (no. 37) ... jUfcL ^j*JI ^
OU-^JIj jJjAII (La Jul t y) L$.:.< could not seriously be inter-

preted as figurative usage or even stylistic option. Their presence in this
collection, like that of nos. 24 and 25 (Q. 28: 76 and 2: 171), can only be
justified by a principle of inclusion which took account of passages
requiring textual emendation.

Now, the possibility of error in the text of scripture, whether as
ungrammatical usage (lahn) or lapsus calami (khatd' al-kuttab), was ulti-
mately rejected. 1 In a transparently dogmatic discussion of two celebrated

dicta, one attributed to 'A'isha: t_itSUI J lyk^t ujBOI J^cjj*, and

the other to 'Uthman: \ajJcu. <_i^*JI jU UjjJu V, Suyuti argued from

the authority of Ibn Anbari and Ibn Ashtah, as well as from his own
conviction, that neither report could be true since (a) the eloquence (fasdha)
of Muhammad's contemporaries was well known, (b) other equally well-
known and better accredited traditions demonstrated the care taken by
'A'isha in the preservation and by 'Uthman in the recension of the
Quranic revelation, and (c) such evidence of textual instability as did
exist was neither lahn nor khatd* al-kuttab, but, rather, scriptio defectiva

or otherwise irregular orthography (1 3, -j l^kiJ or variae lectiones

(o'tl tyt.j). z Reference to Arab eloquence, neutralization of refractory
hadiths by other hadiths, and accommodation of textual variants under the
rubric 'seven (canonical) readings' (e.g. i>u*JI (j^Vl ^» <Jj^' jW^O
reflect the procedural devices traditional to the solution of Quranic prob-
lems, and ought by now to be familiar. 3 It may be observed that Suyuji's
denial of lahn in the text of scripture was part of his general discussion of
i'rab, a term whose semantic range included not only grammatical pheno-
mena (akkam), but also clarity (bayan) and euphony (astodt wa-hihun).*

1 Such is the substance of the first five chapters of Ibn Qutayba's Tawil mushkil
al-Qur'dn (10-75), in which the document of revelation was defended from every kind of
assault upon its linguistic and literary excellence, e.g. lapsus calami, grammatical error,
syntactic inconcinnity, semantic contradiction, and stylistic inconsistency; the author's
postulates were those whose elaboration has been described here: all gird' at were equally
revealed, poetry could be adduced to demonstrate analogous constructions (36-40),
corruption by foreigners (abna al-ajam) was neutralized by the reliable witness of trans-
mission from, companions of the prophet (41-2), etc. * Itqdn ii, 269-77.

3 See above, III pp. 93-9; and 178-82, 202-8, respectively.

4 Itqdn, now' 41: ii, 260-80; see above, p. 155; and III p. 104-11 for references to
Itqdn iv, 172-3, ii, 3.

222

QURANIC STUDIES

It is also in that chapter that locutions permitting any one of the three
i'rab vowels were listed. 1 The possibility of serious linguistic aberration
in scripture was very restricted indeed.
The locus classicus was always Q. 20: 63 jlj^LJ jIIa jl. Suyuti's

synthesis includes the entire spectrum of textual treatment, as well as a
conventional notice of dialectal usage: (a) in some dialects, e.g. of
B. Kinana and B. Harith, the dual is expressed by alif in all three cases;
(A) a pronoun anticipating the subsequent proposition {damir al-sha'n)

has been omitted, i.e. jl^LJ jljkA ( c ) in addition to the ellipsis

assumed in (b) a further omission of an inchoative of which sahiran is

the predicate, i.e. jl^L, L4J jl 1st. Jul ; (d) inna in this locution signifies

'yes/surely/indeed' and thus does not require casus obliquus; (e) ha is
here an anticipatory pronoun {damir al-qissd), and the remainder of the

proposition an independent predication, i.e. jl^LJ jli l$j|. 2 At least

two other kinds of solution had been earlier articulated. In addition to
citing the dialectal usage (here of B. Balharith b. Ka'b), Ibn Qutayba
recorded appropriate variants {tnasahif) from two 'companion' codices,

namely Ubayy jl_^L. ^1 jli jl and Ibn Masud jl^L jU* jl or

jl^*.UJ jiOA jl or ji^L j|jjt> jl. 3 The exegetical character of such

ascription has been noted. 4 Of equal interest was the view, attributed by

Ml wt

Ibn Qutayba to 'Asim Jahdari that one wrote jljt* jj and read jl,
thus providing explicit support for the TJthman tradition on lahn in the
text of scripture, i.e. ; -Oil «u»j jLic Jja) oe.\ Jei\ ^ Jjy LJlj

•OU. —yj AjLL JuU'ti Igi^jti ^->yi\ <U-«2x«j U?J jlJ t£j!. 5

Such was the solution proposed also by Abu 'Ubayda in slightly different
terms: J j^a&i? Oj^ji U5" jlo* i-^Tj Ja^Ul <J j! _^LJ ^j^a j]
i_>!>*» JiiJUlj <_ifc£JI. 6 All of the 'standard' orthographic deviations, not
only Q. 20: 63, but also 2: 177 (wal-sabirin:sabiruri), 4: 162 {wal-
muqimin:muqimun), and 5: 69 (wal-sabi'un:sdbi'in), could be and often
were emended in this way. 7 Other problems, semantic rather than gram-
matical, might be exposed to the same treatment, e.g. Q. 56: 29 {wa-taTin
mandudin for wa-falhin). 6

1 See above, III p. 108, referring to Itqdn ii, 277-80.

2 Itqdn ii, 273-4.

3 Ta'rml, 36-7; see Jeffery, Materials, 146 and 60, respectively.

4 See above, pp. 203-7. 5 Ibn Qutayba, Ta'iml, 37; see GdQ iii, 4-5.

6 Majaz al-Qur'dn ii, 21-3 ; see Wansbrough, 'Periphrastic exegesis', 264.

7 e.g. Ibn Qutayba, Ta'tvll, 37; see above, III p. 108.

* Goldziher, Richtungen, 36 ; Zamakhshari adduced a scriptural analogy from Q. 50: 10,
Kashshaf iv, 461 ad loc.

PRINCIPLES OF EXEGESIS

223

\ The masoretic principle of 3VD : "'Ij? was perpetuated in the X^5~ :
(t^jlx^: otl ^i) of the Muslim exegetes. Its apphcation in the Quranic masorah
may, in my opinion, be ascribed to caique rather than to inherent necessity.
The methodological significance of such devices as ketib.qere and al tiqra
is generally seen to be evidence of a fixed, immutable text. 1 For the text
of Muslim scripture standard deviations were anyway accommodated by
the system of 'canonical' readings, within which a distinction between
kitba and lafz merely symbolized an option (ikhtiydr). 2 It will, I think,
be conceded that formulation of the Biblical masorah included a period
in which emendation to a still fluid text was not only possible, but in fact
took place. The activities attributed specifically to the post-Talmudic
masoretes exhibit only the final expression of a long and complex process,
to which elements of the masorah like tiqqunei soferim and miqra soferim
attest. 3 For the Quranic masorah the concept of emendation is explicit
in the principle of majdz as employed by Abu 'Ubayda: for the solecism in
Q. 20 : 63 the qere (lafz) was expressly 'correct' (saviab). But emendation
was not by any means limited to, perhaps not even primarily concerned
with, irregularities of grammar and syntax. Exegetical, often dogmatic,
ends were equally served. One example, ad Q. 37: 12, has been noticed; 4
another was the alteration of alldhu to allaha in Q. 4 : 1 64, producing 'Moses
spoke to God' rather than 'God spoke to Moses', a change reflecting
Mu'tazili circumscription of the notion of God's speech. 5 However, very
few of Abu 'Ubayda's sixty shawahid required to be emended for reasons
of dogma or doctrine. 6

Masoretic majdz, like the grammarians' principle of taqdir, had been
formulated primarily with a view to obviating the angularities of scriptural
syntax. The norm against which such were measured was often provided
by scripture itself, but also and with increasing frequency by the data
of profane literature. These might consist either of real, or contrived,
examples of the urns loquendi (introduced by qawluka, ka-qawlika), or
of loci probantes from Arabic poetry. 7 In support of the dialectal origins
of hddhdni in Q. 20: 63, for instance, Ibn Qutayba adduced two lines of

1 e.g. Barr, Comparative Philology, 45-6, 214-17; cf. Wiirthwein, Text, 12-22, 71-5.
1 See above, pp. 205-7; and cf. Rabin's reference to the 'limited variability' of the text
of scripture attested also in Qumran, cited Gerhardsson, Memory, 37 n. 2.

3 See Barr, op. cit., 210-21; Gerhardsson, Memory, 33-42, 43-55; and Gertner,
'The Masorah and the Levites', 266-70.

4 See above, p. 206.

5 See above, I pp. 34-6, II pp. 81-4; Goldziher, Richtutigen, 174-5: a more radical
solution to the dogmatic problem which did not require textual emendation was deriva-
tion of kallama in this passage from kalm (wound); cf. Zamakhshari, Kashshdf i, 590-1 ;
but also Ibn Qutayba, Tamil, 82.

6 Wansbrough, 'Periphrastic exegesis', 256: three instances only, to which may be
added, under no. 30, Q. 79: 30; see above, p. 220.

7 See above, pp. 211-12, 216-18; and Wild, Das Kitdb al-'Ain, 44 n. 16.

I

224 QURANIC STUDIES

verse (anonymous) demonstrating the expression of both casus rectus
and obliquus (dual) with alif. 1 The eventual opposition, hardly unexpected,
to that technique, 2 found unequivocal expression in the observation of

Ibn Munayyir to Zamakhshari's commentary ad Q. 6: 137: U^-yl ^^Jj
ecljSJU jLj^JI Jicly £s*^oj Jj A&\yi oe.lj2JI ^^syaj' (Our purpose

is not correction of scripture by reference to Arabic grammar, but, rather,
correction/establishment of Arabic grammar by reference to scripture). 3
It need hardly be remarked, in view of the role played by 'Arabian
eloquence' in Quranic theology, that this admonition remained a very
dead letter indeed. Lip-service to an ideal had, however, been expressed. 4
Another exclusively interpretative element of the Quranic masorah, and
hence one which did not always require textual emendation, concerned
the problem of juncture. An example was the question of pause after

kadhalika in the messenger formula dly <Jlj dU.i5r_(J*JlS/JU) (Q. 19: 9,
19: 21, 51: 30).* The alternatives were to read with pause making the
particle disjunctive, or without pause acknowledging the formulaic locution,
i.e. nVP "J8K TO. Zamakhshari appears to have preferred the former <£)Sjf
JUL (^aJ jt dJj Jlii fj&l jJ jjJw»j dJUii" jJl\ £jj «JtS3l. 6

The term for disjunctive syntax (Zamakhshari employed ibtada) was
usually isti'naf, found already in Muqatil's Tafstr ad Q. 3 : 7, for which
verse of course the argument was doctrinal and not at all syntactical. 7
In his chapter on juncture (al-mawsul lafzan wal-mafsul ma nan) SuyutI
provided several illustrations of that kind of theological grammar. 8 For
Q. 7: 189-90 the charge of polytheism against Adam and Eve was made
quite explicit by context (siydq). But Adam was a prophet (nabi') and
therefore immune from sin {ma' sum). A solution to the dilemma was found
in the change of pronoun from dual to plural, making the referent of

OjS^jLs U* <UI I ,J Uii not the inhabitants of Eden but the pagans of Mecca.
SuyutI called the trope thus isolated an example of 'transition and digres-
sion' (al-takhallus wal-istitrad), 9 ignoring conveniently the distinct likeli-
hood that the form yushrikun reflected the pattern of verse segmentation
at that point of the sura. On the other hand, it may be argued that not merely

1 Ibn Qutayba, Ta'tml, 36. 2 e.g. SuyutI, Itqan ii, 55.

3 Ibn Munayyir, Intifdf, on the margin of Kashshdf ii, 69-70; cited also Goldziher,
Richtungen, 50 n. 3; Ullmann, Ragazpoesie, 222-3 n. 10; see above, p. 168.

4 See above, II pp. 78-81. s See above, I pp. 12-13.

6 Kashshdf iii, 6 ad Q. 19: 9.

7 See above, p. 152; appearance of the technical term isti'naf in Muqatil, Tafsir,
MS H. Hiisntt 17, 36 1 , may well be a consequence of later redaction, cf. above, pp. 143-4.

8 Itqan, nam' 29: i, 252-4, an appendix to his discussion of pausal phenomena {aU
ivaqfieal-ibtidd*): naw' 28: i, 230-51.

9 See Mehren, Rhetorik, 130, 145.

PRINCIPLES OF EXEGESIS

225

s the plural yushrikun but the entire locution exhibits a formula of period-
ization, and need not therefore be semantically related to the preceding
Adam tradition. 1

Easily the most celebrated of such passages was Q. 106: 1 fjt^J \S^S^,

for which the function of lam as inchoative provoked some very imagina-
tive grammatical theory. 2 To those disatisfled with exegesis, textual
emendation was possible: Suras 105 and 106 were read as a syntactical
unit, for which of course support could be found in a 'companion* codex
(i.e. Ubayy). 3 But however the two suras were written (generally with the
disjunctive basmala) they were understood as a narrative unit attesting to
God's benevolence towards Quraysh. 4 Problems of the sort found at Q.
106: 1 have also been noted in the transmission history of the Biblical text,
e.g. D'lOBOS vs. BVO 103 in Proverbs 26: 23, described by Barr as a
'graphic disturbance'. 5 Ellipses and anacolutha in the Quranic text often
exhibit illogical juncture rather than defective syntax: for example,
Abu 'Ubayda's inclusion under ellipsis (no. 3) of Q. 39: 73, in which the
second dependent clause of the hatta idha construction is the following
verse, introduced by the coordinate ( !) conjunction wow. 6

Abu 'Ubayda's contribution to the Quranic masorah consisted primarily
in the rationalization of solecism, but also in the resolution of synthesis
and the alleviation of ellipsis. The periphrastic technique, later designated
taqdir, was also applied to the resolution of trope (a kind of inverse meta-
phor, as for Q. 12: 82) and to the elimination of anthropomorphism. 7 For
the history of the text of scripture one question in particular requires at
least to be articulated, if not answered: was the principle of majdzjtaqdir
understood as emendation or as exegesis? Now, copies of the Qur'an,
whether manuscript or printed, do not contain masoretic material, and
may for that very reason be misleading: the impression is unmistakably
that of a single ne varietur text. But even the most cursory examination of
the masoretic literature must dispel that impression as illusory. The
chronology of that literature indicates a period of approximately two

1 See above, III pp. 113, 115, 117.

2 Analysed in Birkeland, The Lord guideth, 102-30.

3 Suyuji, Itqdn i, 186; Jeffery, Materials, 179; sectarian appropriation of that view is
significant only as an expression of opposition to the "Uthmanic recension, cf. GdQ ii,
33 n. 4, 96 n. 3.

4 e.g. Ibn Qutayba, Ta'wil, 319-30; and Birkeland's interpretation, op. cit. 122-30:
'God's intervention in history as the first stage of Muhammad's prophetical experience'.

s Comparative Philology, 219; cf. also Gertner, 'The Masorah and the Levites', 247
for Exodus 23 : 2.

6 See above, III pp. 114-15; Wansbrough, 'Periphrastic exegesis', 248, 255; Brockel-
mann's explanation of that phenomenon, G VG ii, 669, as 'die Neigung des semitischen
Sprachgeistes zu einfacher Satzbildung eine komplizierte Periode wieder in einfache For-
men (zu zerlegen)' cannot, in my opinion, be supported by Quranic examples.

7 Wansbrough, op. cit. 254-9.

4339C75 I

226

QURANIC STUDIES

centuries during which the text of scripture was anything but stable.
The major obstacle to reconstruction of the transmission history of that
period is the character of the so-called 'variant readings', which is pre-
dominantly exegetical: in other words all, or almost all, are clearly inter-
pretations of the canon. Within the masorah their relative merits were
assessed according to their respective degrees of attestation. The apparatus
criticus resulting from such accumulation of essentially uniform materials
exhibits a series of conventional modifications of one text. At the same
time it conceals, and may have been meant to conceal (by means of the
'serial revelation' argument), 1 the fundamental lines of cleavage that I
have called 'variant traditions'. But indirect notice of these may be thought
to underlie the masoretic categories of wujuh, nazair, mushtabihat, and
majaz. Collation of parallel and divergent contexts, deductions from analo-
gical and identical phraseology, rationalization of grammatical and syn-
tactic irregularities: all of these presuppose a consciousness of recurring
formulae and schemata. According to the theoretical postulates of the
grammarians, a restored text (muqaddar) was at least as valid as its original
articulation (malfuz). 2 Preservation and transmission of those restorations
were for the masoretes tasks equal in importance to recording the canoni-
cal text. 3 The 'limited variability' of the canon was a fact, the 'Uthmanic
recension an article of faith As such, exegesis and emendation might be
seen as two names for the same process, and not as mutually exclusive.
Like the halakhic concept of preference (tarjih), the masoretic principle of
option (ikktiydr) contained an arbitrary element which never had to con-
form to a single, fixed text.

And yet one feature of masoretic exegesis, as contrasted with the
haggadic and halakhic types, was the practice of adducing and (usually)
of commenting on the entire text of scripture. A concern for the integrity
of the text and for the structural relevance to one another of its parts was
characteristic even of purely lexical analysis. For the analogical restoration
of grammar and syntax it was indispensable. The 'framework' of masoretic
exegesis was thus neither the narratio (haggadic) nor the juridical/doctrinal
dispute (halakhic), but the canonical text itself, perceived as a unitary
document. Within that framework 'explicative elements' like variae
lectiones, lexical and grammatical explanation, analogy, periphrasis, and
loci probantes from poetry were typical. 4 All other elements may be re-
garded as intrusive, and none more than authentication of the masorah
(i.e. as apparatus criticus) by reference to traditions and to the circum-
stances of revelation. That such should, none the less, figure in this genre
of exegetical literature demonstrates, in my opinion, the chronological
priority of the halakhah, in support of which a basis of traditional authority

1 See above, I pp. 36-8. 2 Wansbrough, op. cit., 257, 264-5.

3 See above, pp. 204-5. 4 See above, pp. 120-1.

PRINCIPLES OF EXEGESIS

227

(i.e. masorah) had to be formulated. In this context Gertner's description
of the separate origins and later juxtaposition/identification of Masorah
and Masoreth is of some interest for the analogous process in Muslim
exegetical literature, namely, the evaluation of rationally deduced (!)
'readings' in terms of their respective degrees of attestation (the evolution
diraya ~>riwdya). 1 It was only after the articulation of law as divinely
decreed that a scriptural canon was established, the result primarily of
polemical pressure. 2 Once stabilized, the document of revelation was no
longer exclusively the 'word of God' but also, and equally important, a
monument of the national literature. In that capacity its service to the
community, and to the cause of polemic, was unlimited.

5. RHETORIC AND ALLEGORY IN EXEGESIS

Acknowledgement of metaphor in the language of scripture could be as
much an expression of piety as of aesthetic appreciation. Elimination of
anthropomorphic imagery predicated of God was the primary form of
that piety, practised by both Quranic and Biblical exegetes. 3 But whatever
the original motive, exegetical speculation was ultimately coupled with
recognition of scripture as the articulation of literary forms and related to
an attested rhetorical tradition. For example, both the obvious and quite
unnecessary insertion (majaz) of ahl into Q. 12: 82 and the not so obvious
but equally unnecessary insertion (taqdir) of amr into Q. 89: 22 4 were
adduced by Sharif Murtada (d. 436/1044) as illustration of the particular
capacity for figurative expression (Tnajazdi) of the Arabic language: 5 jjitl

tJjtftJI L-,(l2: 82) kyjJI jLlj (89: 22) duj tUj ,JUJ A)j» 6Vj jlya^-^lj

Stress in that passage lay on the qualities of conciseness iijaz),

brevity (ikhtisar), and ellipsis (hadhf) inherent in Arabic, thought for that
reason inter alia to be superior to other languages. The shift of emphasis by
means of which those and other phenomena of the scriptural style became
points of departure for the elaboration of an extended corpus of literary
theory is the subject of this, the concluding, section of these studies.
The standard apologia for figurative usage in scripture would seem to

5 Gertner, op. cit., esp. 255-9. * See above, pp. 158-63, 194-5, aoi-a.

3 See Wansbrough, 'Periphrastic exegesis', 259 (Ibn IJazm), 262, 264 (Saadya), 266
(Maimonides) ; cf. Goldziher, Richtungen, 116 nn. 2, 3, 7.

4 Wansbrough, op. cit. 254-5, 257, 259.

5 GAL i, 404-5, Suppl. I, 704-6; Amali ii (Takmila), 309-1 1.

aa8 QURANIC STUDIES

presuppose an expressly formulated opposition to such, and occasionally
indeed, that opposition was identified with scholars of the Zahiri, Maliki,
or Hanbali schools. 1 In fact, even the alleged opponents of Quranic
majdz were constrained, when confronted by anthropomorphisms at
least, to resort to exegetical devices like taqdir for the resolution of meta-
phor. 2 The apologia was none the less articulated, in a form both simple

and effective, by Ibn Qutayba: 3 p^Ili j LjmJU jT^JI ,J* o^JLcUaJi Ufj

^ |Ju>j(i2:82) JU ^ ^13(18: 77) Jujj'il jlJbJI

cj jyj civ ijuu t^r jjrf oir-^u d\j~*i\ jj- ji v-*^

^Jl t^-jj J^l o^l C^Jj ij*JJ\ oJU>j JJUI- The truism

that for the mimetic function of speech metaphor was indispensable
symbolized formal and collective recognition of an exegetical factor
common both to halakhic dispute (ikhtilaf) and to masoretic emendation
(majdz), namely that language could not be construed as having merely or
exclusively an immediate (verifiable and quantifiable) relation to the data
of experience it purportedly described. A good deal of halakhic exegesis
turned upon that very point: e.g. derivation of a series of graded punish-
ments for muharibun from the syntactic sequence of Q. 5 : 33, 4 or the (per-
haps) extreme argument according to which a blind husband could be
excluded from the provisions of Numbers 5 : 13. 5 For the Quranic masorah,
Abu 'Ubayda's majaz, when not directed to flagrant examples of gram-
matical irregularity, was applied to idiom and conventionally ambiguous
usage, rather than to the analysis of metaphor as consciously formulated
imagery. 6 Ibn Qutayba's monograph on the style of scripture exhibits the
transitional employment of majaz: from an interpretative device to an
aesthetic category.

The earlier sections of that work are concerned primarily to refute
allegations of solecism (lahn) and contradiction (tandqud) in the text of
scripture, and belong thus almost entirely to the masoretic tradition
(i.e. chapters I-V) as exemplified by Farra' and Abu 'Ubayda. 7 It is in the
middle sections (i.e. chapters VI-XII) that the author treated the pheno-
mena of figurative language, after a general discussion, under six headings : 8
metaphor, inversion, ellipsis, repetition and pleonasm, metonymy and

1 e.g. Suyflti, Itqan iii, 109.

2 e.g. Ibn Hazni, see Goldziher, Zdhiriten, 164-8.

3 Ibn Qutayba, Ta'wil, 99. « See above, pp. 187-8.
s Gertner, 'Terms', 20.

6 See above, p. 230.

7 Ta'wil, 10-75; see above, p. 331 n. I.

* Ta'wil, 76-229 subdivided: general (76-101), metaphor (102-41), inversion (142-61),
ellipsis (162-79), repetition and pleonasm (180-98), metonymy and allusion (199-212),
idiom (213-29).

PRINCIPLES OF EXEGESIS

229

allusion, idiom. At only one point (though admittedly conceivable in a
number of other similar contexts) did he employ the locution wa-majazuhu
in the manner of Abu 'Ubayda (i.e. 'and its restoration is' or 'and it ought
to read'), namely, for the not very problematic expression ^50

Q. 55 : 31, interpreted as 'And we will attend to you (scil. after long neglect
and delay)'. 1 That for Ibn Qutayba majdz did not only signify metaphor
(isti'dra) is clear not merely from the organization of his loci probantes
but from his express declaration: ^ij jlseJI jS^\ ojUz^l oL IjlJj
jus. 2 His illustrations of isti'dra range from examples of genuinely tropical
usage, like the expression u, ^ i_i£5o ^ in Q. 68 : 42, in which
'shank' is a metaphor for resolution/energy (ye. JLz y* oJui ye. ^

i'-l£J1 ^}>yi J {JjUJ! c*j^£~\a aSL.), 3 across the onomatopoeic ejaculation
uff in Q. 17: 23, explained as an instance of synaesthetic formation
(dAli j~c-j .iLj jf Cx* < ^W Lft diiaHi* <_)*>!>),♦ to the

standard example of ellipsis in Q. 12: 82. s In view of the ubiquity
as locus probons of Q. 12: 82 in exegetical literature, it is worth re-
cording that this and other instances of ellipsis (hadhf) were not, when
unaccompanied by a change in i'rdb, considered to qualify as figurative
usage by the major theorist of Arabic rhetoric, 'Abd al-Qahir JurjanI
(d. 471/1078). 6 That argument, applied also to pleonasm (ziydda), 1 drew
attention to the essentially stylistic function of both phenomena by stressing
the intentional dislocation of the entire utterance. Ellipsis and pleonasm
were thus regarded as tropes and required to be appreciated, rather than
merely clarified or emended. Ibn Qutayba's inclusion of Q. 12: 82 under
metaphor (isti'dra) might be thought evidence of a similar, if somewhat
less sophisticated, point of view. The fact that Q. 12: 82, together with
2: 177 and 47: 13, was adduced also in his section on abbreviation/ellipsis
(ihhtkdrjhadhf) may betoken some indecision about that construction. 8
But the position of Ibn Qutayba in the evolution of rhetorical exegesis is,
in my opinion, transitional, though his precise description of the nature of

the ellipsis in Q. 12: 82 corresponds to that of Jurjani, i.e. ^Jj&cS jf
«J J*«JI Jjc>Jj juIL. -Ul oUkJI pJJj <Jl^J1. It was that connection
of ellipsis and pleonasm with syntactic function (i'rdb) which was
preserved in the exegetical tradition. 9

• Tamil, 77.

2 Ta'ml, 101 ; cf. provisional definition of majdzdt fil-kaldm, 15-16.

3 Ta'vM, 103. 4 Ta'vnl, 111.

5 Ta'wil, 129, applied also to Q. 44: 29 and 47: 4.

6 GAL i, 341-2, Suppl. I, 503-4; Asrdr al-Baldgha, paras. 26/1-3; see Wansbrough,
'Periphrastic exegesis', 255 n. 10. 7 JurjanI, Asrdr, paras. 26/4-10.

* Ta'iml, 162. ' e.g. SuyutI, Itqdn iii, 124-5, citing ZanjanI and Qazwini.

230

QURANIC STUDIES

Ibn Qutayba's second category of majaz includes two kinds of 'inversion'
{maqlub) : one semantic, the other syntactic. The first consists primarily of

locutions per antiphrasin (<ui^ ^ti\ \J^y_ <j\ <_jjUJI ^j) employed

as omen (tatayyurjtafd^ul) particularly in onomastica, but also as hyperbole
(mubdlagha) and ridicule (istihza). An example of hyperbolic usage, or
more correctly perhaps of litotes, was zann (conjecture) for yaqin (cer-
tainty) in contexts of eschatological reference, in which of course there
could be no question of 'doubt', e.g. Q. 2: 249, 18: 53, 69: 20. 1 Syntactic
inversion, on the other hand, was for Ibn Qutayba hyperbaton, as in
Q. 3 : 40 'And old age has overtaken me' for 'I have attained old age'. 2
Although he appears to have denied presence in the Qur'an of hysteron
proteron (maqlub 'aid 'l-ghalat), 3 the inclusion of recognized 'problem
passages' like Q. 2: 171 and 28: 76, as well as 18: 1-2, must be interpreted
as tacit acknowledgement of such. 4 None of the three exhibits the rhetorical
embellishment illustrated by Q. 3 : 40.

Similarly, Ibn Qutayba's eight kinds of ellipsis include several examples
of the rhetorically effective omission of an apodosis, e.g. in a hypothetical
construction (Q. 13: 31), and in an oath (Q. 50: i). 5 But ellipsis is also
represented by zeugma, e.g. Q. 10: 71 and 17: 23, and also by sheer care-
lessness, e.g. Q. 38: 32 and 97: i. 6 Mention there of the synthetic
construction in Q. 83: 3 (wazanuhum for wazanu lahum) can only be
understood as a survival from the masoretic tradition. 7

In his treatment of repetition and pleonasm (takrar uoa-ziyddd) the
author distinguished on the one hand repetition of narrative passages,
duly related to the doctrine of 'serial' revelation (munajjamanlnujumari),*
and on the other, the verbatim repetition of specific locutions, as in Suras
55 and 109. 9 Reason for both, as for repetition of single words in verses
like Q. 2: 196 and 7: 12, was emphasis and drill (tazokid wa-ifham), a
view of that particular Quranic phenomenon which has informed all sub-
sequent scholarship. 10 Unlike Jurjani, Ibn Qutayba applied the term ziydda
exclusively to (in his opinion) otiose elements like the particle bi in Q. 96:
1, 11 but also the word wajh (face) in Q. 2: 115, 6: 52, 28: 88, and 76: 9,

1 Ta'wil, 142-4. * Ta'wil, 149.

3 Ta'wil, 154.

* Ta'wil, 153-8; cf. Wansbrough, 'Periphrastic exegesis', 251-2 (nos. 24 and 25), and
above, p. 220; Zamakhshari, Kashshaf \\, 702 ad Q. 18: 1-2.

5 Ta'wil, 165, 173, respectively.

6 Ta'wil, 164, 167, and 174; the pronominal reference in Q. 97: 1 was in fact a matter
of doctrinal significance, see above, II p. 62.

7 Cf. Wansbrough, op. cit., 253 (no. 31) and 256.

8 See above, I pp. 36-8.

» Ta'wil, 180-2; see above, I pp. 25-6.

10 See above, III pp. m-12.

11 Thus also Abu 'Ubayda, see Wansbrough, op. cit., 253 (no. 33) and 257.

PRINCIPLES OF EXEGESIS

231

v exhibiting of course neither grammatical nor rhetorical, but rather doc-
trinal concern for the anthropomorphic attribute. 1

Allusion {ta'rid) was terminologically distinguished from metonymy
{kinaya), though Ibn Qutayba's analysis of the two was, rightly, synoptic. 1
The latter includes the kunya itself (onomastic), the kind of allusion com-
monly resolved by haggadic ta'yin, e.g.fulanan in Q. 25: 28, 3 and finally,
the generic application of the definite article, e.g. al-zdlim in Q. 25 : 27
or al-kafir in Q. 78: 40. The last represented a rhetorical elaboration of the
halakhic khusus/'umum argumentation. 4 For Ibn Qutayba, ta'rid itself
meant the kind of euphemism exhibited in Q. 18 : 73 or the circumspection
of Q. 34: 24. In both passages harshness was alleviated by recourse to
circumlocution, but not to the extent of suppressing altogether the facts
of 'forgetting/forgetfulness' and 'error/sin', respectively. The opposite of

ta'rid in that sense was tasrih; its synonym was tawriya (^a ^ <s]j)- s

Of some interest for the development of rhetorical exegesis is comparison
of Ibn Qutayba's treatment of Q. 34: 24 with Abu 'Ubayda's grammatical
cum dogmatic 'restoration'. 6

It is in the section dealing with idiomatic expressions (entitled iaJUt*

aluw Jaill I y>Ui) that Ibn Qutayba adhered most closely to the masoretic

tradition, treating in turn problems of morphology, tempora, juncture,
number, and specification. 7 Three other elements belong more properly
to rhetorical analysis: (a) imprecation as divine utterance (e.g. qa.talah.umu
'llah in Q. 9: 30) was interpreted as hypothetical; (b) rhetorical questions
were analysed as signifying affirmation (taqrir, as in Q. 20: 17), wonder
(ta'ajjub), or reproach (tawbikh) ; (c) prohibitives/imperatives might convey
threat (tahdid), admonition {ta'dib), even exemption (ibdka, as in Q. 62:
io). 8 Thus, majdz in the work of Ibn Qutayba, as for Saadya, might be
understood to include not only trope, but also idiom and popular usage. 9
That the latter should be subsumed under the general rubric of rhetorical
device may, in my view, be attributed to the dominant role of scriptural
exegesis in the elaboration of Arabic literary theory. As in philology, so
in rhetoric the tyranny of lingua sacra was not merely felt, but found
expression as a criterion of excellence. However the dogma of i'jaz

1 Cf. Saadya ad Psalm 88: 15, in Wansbrough, op. cit. 264.

* Ta'zvil, 199-204; that pattern of exposition was characteristic also of the later theor-
ists, see von Grunebaum, Tenth-century Document, 38-9 n. 297.

3 See above, pp. 135-6. * See above, pp. 169-70, 191.

5 Ta'zvil, 210, but also the entire section, 204-12 ; thus considerably earlier than Zamakh-
shari, pace Bonebakker, Tawriya, 27-8.

6 Wansbrough, op. cit. 256-7.

7 Ta'teil, 213-29: morphology (228-9), tempora (227-8), juncture (226-7), number
(218-26), specification (217-18).

8 TcCisAl, 213-15, 215-16, 216-17, respectively.

9 Wansbrough, op. cit. 265-6: references in nn. 79-80.

232

QURANIC STUDIES

al-Qur'an was interpreted (theologically or rhetorically or both), compari-
son of the profane and sacred styles was inevitable. A perfect illustration of
the double standard applied to such comparison was the observation of
BaqillanI about a line from Imru' 'l-Qays: that JL43 <r>j^- ij* U
ought to have read ..... l#*~o U, and that even poetic licence could
hardly justify interpretation of the pronoun ma as feminine. 1 Now, in the
light of such examples of wayward concord as Abu 'Ubayda's nos. 34 and 35,
that stricture must appear harsh if not perverse. 2 But Baqillam's contribu-
tion to the science of rhetoric was marginal indeed : it was his merit as a
theologian to formulate the i'jaz argument in terms borrowed from the
works of contemporary rhetorists, but not without the over-simplification
inherent in synthesis. 3

With its application to the scriptural style, rhetorical terminology exhi-
bited evidence both of mutation and of proliferation. Much of that was the
consequence of seeking, and findings in the text of scripture at least one
example of every figure known to the profane tradition. A trace of embar-
rassment, not quite concealed even in the assertive and confident approach
of BaqillanI, led to increasing terminological differentiation in order to
prove the divine origin of all rhetorical device. A minor but none the less
significant illustration of that process was the evolution of the trope known
as madhhab kalami: from conceit to the syllogistic formulation called
enthymeme. 4 The figure itself appears to have been originally the parono-
mastic epigram, of which a most artful example was composed and included
in his rhetorical treatise by Ibn al-Mu'tazz (d. 296/908) : s

JUL tjfi 1* J ^ jjj

Now, to locate in the document of Muslim revelation so cunning an artifice
as that would have required considerable ingenuity, and it is hardly
surprising that the earliest theorists, e.g. Jahiz and Ibn al-Mu'tazz himself,
denied its presence there. 6 When eventually the madhhab kalami was
discovered to be of scriptural origin, the figure had altered quite beyond
recognition, the work of two late scholastic theorists: Ibn Abl '1-Isba'
(d. 654/1256) and Khatib Qazwini (d. 738/1338). The locus classicus was
Q. 21 : 22 Ij J^Ji AJi\*)} \ igJ? _jJ (If there were in them, scil. heaven

1 BaqillanI, I'jaz al-Qur'dn, 161; trans, von Grunebaum, Tenth-century Document, 63.

2 Wansbrough, op. cit. 253.

3 See von Grunebaum, Tenth-century Document, 6 n. 43; id., Kritik, 87-100: the
'missing work' (97) is Ibn Wahb, Burhdn, cf. BSOAS xxxiii (1970) 616.

4 See Wansbrough, 'Note', 55-63.

s GAL i, 8i ; Kitab al-badi', 56; trans. Wansbrough, 'Note', 59.
6 Kitab al-badi', 53.

PRINCIPLES OF EXEGESIS

333

and earth, gods other than God, both would have perished), that verse
being interpreted as an argument for a single author of creation. The
transition from epigram to dialectic (with the suppressed middle term
characteristic of the philosophers' enthymeme) can only be explained by
reference to the earlier inclusion under madhhab kalami of parodistic com-
positions ridiculing the language of philosophers and theologians. 1 In the
search for Quranic shawahid the element of parody had naturally been
ignored, or forgotten, and each example exhibited, at least vaguely, a kind
of apodictic syllogism : e.g. since the repetition of divine creation is easier
than creation itself, it (scil. resurrection) is ipso facto possible (Q. 30: 27);
the moon may vanish, but God does not vanish and therefore the moon
cannot be God (Q. 6: 76); you are punished, but the sons (of God) are
not punished, therefore you are not sons of God (Q. 5 : 1 8). 2 For Ibn
Abi '1-Isba r , Q. 21 : 104 and all related assertions of the resurrection were
employed to illustrate the madhhab kalami, the arguments an elaboration
of the type employed by QazwinI for Q. 30: 27. 3 It seems unlikely that
madhhab kalami would, without the challenge offered by scripture, have
evolved much beyond its employment as caricature of technical jargon,
e.g. 4

or

Application of the term to so serious a subject as arguments in support
of monotheism (the theme common to all of the scriptural shawahid)
might be thought to require a very sharp divergence of the profane and
sacred rhetorical traditions. 5 It could even be argued that description of
the phenomenon as exegetical appropriation of a profane terminus technicus
is facile and simplistic. 6

Rather more complex than the mutation of madhhab kalami was the
rhetorical-exegetical development of the figure originally called tafstr. -7
While for madhhab kalami retention of one name for three separate pheno-
mena might justify a hypothesis of adaptation, the evolution of tafstr
into laff wa-nashr {inter alia: versus rapportati) involved changes in form,
content, and nomenclature. The specifically exegetical residue from that
compound process consisted of two Quranic constructions: the first

1 Wansbrough, 'Note', 58-62.

2 Qazwini, Iddh, in Shuruh al-talkhis iv, 369-70.

3 Ibn Abi '1-Isba', BadV al-Qur'dn, 37-42; Suyuti, Itqdn iv, 52-5.

4 Subki, 'Arus, in Shuruh al-talkhif iv, 372-3.

5 Cf. Goldziher, Zdhhiten, 133.

6 Cf. Wansbrough, 'Qur'anic exegesis', 469-70: it seems to me unlikely that Ibn Athir's
ma'tid 'l-san'a can have signified more than 'schemata' in general, cf. Heinrichs, Arabische
Dichtung, 91 n. 3.

7 Wansbrough, 'Qur'anic exegesis', 469-85-

234

QURANIC STUDIES

designated mufassal (separate/diffuse) and mujmal (composite), was nothing
more than a proposition containing a gloss in the form of a sub-
ordinate clause. Hence, its original name: tafsir (subnexio). The
difference between the mufassal and mujmal varieties lay in the ratio
of elements in the gloss to those/that of the referent: in the former
there were two or more in each, in the latter two or three in the gloss to
one in the referent. Frequently adduced, and certainly the most graphic

illustration of the mujmal construction was Q. 13 : 12 JjljJI p&ji _>*J
UJsj Up.. 2 Now, the exegetical moment in the tafsir I laff wa-nashr evolu-
tion was not quite the same as that in the development of madhhab kalami.
In the latter the evidence suggested that Quranic loci had at all costs to be
found for every component of rhetorical ornatus (badt); 3 in the former a
genuine problem of scriptural syntax, subtly identified with a trope well
established in profane literature, was lent a kind of rhetorical legitimacy.
One element common to both problems, however, deserves notice, namely
justification of lingua sacra by reference to the data of profane rhetoric.
In practice at least, if not in theory. 4

For a figure conventionally represented by sequences of multiple
imagery, e.g.

the role of mujmal constructions in exegesis remained oddly anomalous,
despite the likelihood that the exegetical laff wa-nashr owed its name, if not
its very existence, to just such constructions. 5 Scholastic elaboration of the
figure produced a number of useful modifications designed to accommodate
an infinitely variable ratio of gloss-elements to referent-elements, e.g. jam',
tafriq, taqsim, and combinations of all three. 6 The sharply defined distinc-
tions between adverbial and relative constructions, and between explicit and
implicit connection of gloss with referent, were thus gradually attenuated. 7
The final synthesis included apposition as well as attribution and predica-
tion, so long as either referent or gloss contained at least two elements.
All such phenomena could be covered by the term tafsir, if not always by

1 Wansbrough, op. cit. 478-82. 2 Wansbrough, op. cit. 475.

3 Thus, the work of Ibn Abi '1-Isba' might be described as the consummation of that
begun by Ibn al-Mu'tazz.

* See above, p. 234. s Wansbrough, op. cit. 471, 481-3.

6 Set out in Qazwini, Talkhif, in Shuruh al-talkhis iv, 329-47; Mehren, Rhetorik,
108-11. 7 Wansbrough, op. cit. 477-80, 483-4.

PRINCIPLES OF EXEGESIS

235

laff wa-nashr. That 'tafsir' was as much the product of concern for rhetoric
as for 'interpretation' in general seems certain. 1 The polarity represented
by tafsir .ta'ztnl, obscured in most varieties of exegetical literature, was
for the most part maintained in rhetorical exegesis.

In the Amali of Sharif Murtada, for example, ta'wil is employed
throughout for the interpretation of scripture, of tradition (hadith), and of
historical reports (akhbar), while tafsir designates, at least in the Supple-
ment (takmila), the interpretation of poetry. The essentially literary
character of that work is evident even in its external structure (dmdlij
majdlis), within which the analysis of poetry was skilfully and felicitously
blended with that of the three basic forms of Arabic prose. 2 Despite pre-
dilection for Mu'tazili authorities and reasoning, the author's generosity
in matters of dogmatic controversy is ubiquitously apparent, and quite
explicit in his observations on the question of juncture in Q. 3: 7: even
if the rasikhun were syntactically separated from allak, and such was by no
means necessary, it was essential to recall that their exegesis could in many
instances be no more than conjectural. 3 Murtada's method was to adduce
all possible aspects (zoujuh) equally weighted and documented, an example
of which may be seen in his five proposals for reconstruction of the proble-
matic syntax of Q. 2: 171. 4 Only one of these required acknowledgement
of the equivalence faihmaful (as in Abu 'Ubayda), a typically masoretic
device; the others reflected solutions of common sense based upon very
reasonable, and obvious, suggestions of ellipsis, e.g. 'The example of him
who admonishes (wa'iz) the disbelievers . . .'. s Similarly, ad Q. 17. 85 the
haggadic tale of a 'rabbinical' test of prophethood was rejected on the
grounds that a question about the Spirit (ruh), had it ever been posed,
could not be a snare and therefore not the occasion which provoked the
Quranic revelation. 6

The conspicuously rationalist approach of Murtada might also be applied
to the logic of scriptural style, e.g. an isolated and somewhat ambiguous

utterance in one of the Shu'ayb traditions : 7 (Q. 7: 89) iy6 <jf LO jj5G Lj
Ujj <U)I ciLi <jf ^1 m reply to the question: could God will sin and/
or disbelief? he distinguished cultic and legal prescriptions ('ibdddt wa-
shar'iyydt) from the elements of belief or dogma (i'tiqdddt), and produced
seven arguments to demonstrate that for one who had professed his faith in
God membership of any confessional community (milla) other than God's

1 See above, pp. 121, 154-6.

2 See Goldziher, Richtungen, 114-17; id., 'Fachr al-dln al-Razi', 216.

3 Amali i, 439-42: majlis 33; cf. Jassas, above, pp. 151, 154-5.

4 Cf. Wansbrough, 'Periphrastic exegesis', 251-2 (no. 25); and above, pp. 220, 230.

5 Sharif Murtada, Amdlli, 215-19: majlis 15.

6 Amali i, 11-12: majlis 2; see above, pp. 122-7.

7 See above, I pp. 21-2: component VI in version A.

236

QURANIC STUDIES

was impossible. 1 The phrase 'unless our Lord God wills it' could be under-
stood only as acknowledgement of God's infinite mercy, not as allusion to
unpredictable and capricious behaviour. Mu'tazili theology drew its precepts
from the intention, as well as from the formal expression of the theodicy.
But Murtada could also express interest in the parts of speech, e.g. the

particle ka in Q. 42 : 1 1 t-^S, ^^aS'^j^J was not to be interpreted as pleonastic
embellishment (ziydda) to mithl, but as altering the quality of negation in
laysa: from specific to generic, analogous to the relation of ma in to ma. 1
In that argument the operative factor was the function of ka in the entire
phrase, rather than merely as (tautological) proclitic, and the reasoning may
be compared with that of Jurjani for the same construction. 3

Save for a very few isolated vestiges of the masoretic tradition, majaz in
exegetical writings after Ibn Qutayba signified figure or trope. That
scholar's defence of metaphor in the language of revelation found expres-
sion more precise and sophisticated with Jurjani, whose principal concern
was to establish the role of context in figurative usage. His method was to
stress the difference between the primitive/traditional symbolic value of
separate words (e.g. yad as ni'tna, or yad as qudra) and the variable func-
tion of such in extended imagery (e.g. the impossibility of saying/writing:
The 'hand' (as 'benefit'/'power') manifested iteself in the land). 4 Thus
was formulated the antithesis majdz-.haqiqa (tropical: veridical), differ-
entiation of which required attention both to context (ta'lif/nazm) and to
the psychological participation (ta'azowul) of the hearer/reader. 5 The
cardinal point of Jurjam's thesis, however, lay in his insistence that the
language of scripture was neither more nor less than the established lexical
stock of Arabic as habitually employed by speakers of that tongue, and that
the incidence of figurative usage, wrongly denied by some and equally
wrongly exaggerated by others, corresponded to the character of the lan-
guage as a whole, profane or sacred: 6 iiilkJ I <ki^*j jf

^LoVlj tjjbeJIj JJUJIj. The significance of that argument can hardly
be overstated. Assessment of the lingua sacra as partaking of the normal
potential in Arabic for rhetorical embellishment and stylistic variation

1 Amdli i, 402-5 : majlis 30.

2 Amdli ii (Takmila), 311.

5 Jurjani, Asrdr, paras. 26/5-8; Murtada did not, however, adduce the condition of
change in the i'rdb in order to distinguish pleonasm from trope.

4 Asrdr, paras. 21/1-16.

5 Asrdr, paras. 23/3, 23/5, 23/10; Wansbrough, op. cit., 266.

6 Asrdr, paras. 23/12-16, esp. 15.

PRINCIPLES OF EXEGESIS

237

i| was the acknowledged point of departure for the analysis of scripture as
literature. Even the heavy-handed tactics of theologians concerned to
demonstrate the inimitability of the Qur'an or the divine origins of the
Arabic language addressed to Muhammad never quite obscured that basic
premiss. Among the several disciplines competent exegetes were ex-
pected to acquire figured the rhetorical trivium: mcfani, bayan, bad?. 1 In
his final synthesis of the Quranic sciences Suyuti devoted seven chapters
to the components of scriptural rhetoric*

In SuyutTs synoptic survey of the exegetical tradition categorical
distinctions were inevitably effaced and terminological niceties blurred.
The binary opposition majaz : haqiqa was, for example, not demonstrated
but merely asserted. Following what must have been the tradition from
Jurjani, majaz was described as either conceptual (aqli) or formal (lughaun),
exhibited respectively in compound constructions (tarkib) and in individual
words (mufrad). 3 An example of the former was Q. 8 : 2 'When His signs
are recited to them they are increased^! faith', in which the causality
inherent in 'increase' was related to the fact of recitation ; an example of the
latter would be Q. 55 : 27 'The face of your Lord endures', in which 'face'
stood in place of being/essence (dhat). The first example might qualify as
a general illustration of tropical usage, the second only as an exegetical
constant (to eliminate anthropomorphism) in the scriptural lexicon.
Majaz had indeed become, with specific reference to the Qur'an, a vague
and general designation of all phenomena requiring to be understood other
than literally, and finally included most of the textual irregularities noted
in the masoretic tradition, e.g. ellipsis, repetition, concord, and morpho-
logy. 4 But in Suyuti's discussion, a curious and illogical blend of the
material inherited from both Abu 'Ubayda and Jurjani, there is a token
effort to circumscribe the field of majaz by excluding or at least questioning
the inclusion precisely of ellipsis, emphasis, simile, metonymy, chiasmus,
and apostrophe. 5 Trope in scripture remained thus a subject of unresolved
controversy. One refinement in particular deserves notice: a kind of com-
pound majaz (majaz al-majaz) was perceived in verses like Q. 7: 26 'We
have caused to descend upon you raiment', analysed as rainfall producing
flax from which garments could be made. 6 That postulate of divine
causality in three stages exhibits a greater concern for dogma than for
rhetoric, but might be thought to reflect at least roughly Jurjani's very
subtle discussion of the fantastic aetiology (ta'ltl takhytK) amply attested
in profane literature. 7 The application for theological purposes of aesthetic
criteria tended to result in mechanical formulations of the sort produced

1 According to one tradition 15 such; see Suyflft Itqan iv, 185-8.

2 Itqan, anted' 52-8: iii, 109-389.

3 Asrdr, paras. 22/10, 25/1-2; Itqan iii, 109-10.

♦ Itqan iii, 11 1-23. 5 Itqan iii, 124-6.

6 Itqan iii, 127; cf. Isaiah 55: 10-11. 7 Asrdr, paras. 16/1-84.

238

QURANIC STUDIES

in abundance by the later schoolmen. 1 That tendency is perfectly illus-
trated in Suyuti's discussion of ornatus (badi'), in the course of which and
on the authority of Ibn Abi '1-Isba' no less than forty-three separate figures
were found in the text of scripture. 2 The long and complex history of most,
if not all, of those figures would undoubtedly show, as for madhhab kalami
and laff wa-nashr, some very arbitrary procedures of identification and/or
of adaptation. The models, as well as the terminology, were indisputably
profane in origin. The contrary might be asserted, but could not be demon-
strated. Another, related instance has been noted: despite theological
objections to their similarity, description of Quranic verse segmentation
was derived, with very little modification, from the technical vocaculary
pertinent to rhymed prose. 3

The detection and analysis of rhetorical convention in scripture went
some way towards the isolation of typical structures, but not quite so far as
recognition of traditional schemata. 4 In Muslim exegetical literature the
rhetoric of scripture was denned in terms of the particular historical and
psychological relationship between God and His prophet. In Orientalist
scholarship the cynosure was shifted just slightly from there to the relation-
ship obtaining between the prophet and his public, a point of view already
implicit and occasionally explicit in the Muslim tradition. That approach
to scriptural rhetoric is adequately illustrated in the studies of Sister,
Metaphern und Vergleiche im Koran (i93i);and Sabbagh, La metaphor e dans
le Coran (1943). An element common to both is a description of imagery
which could almost be called sociological: the acquisition by one man
of linguistic expressions within a cultural environment whose components
were familiar and verifiable because so widely and well attested. Now, the
examination of available source materials, such as I have attempted in these
studies, would hardly seem to support the assumption of Urerlebnis
exhibited in the analyses either of Sister (e.g. 'Die Natur: Himmel und
Gestirne, Gewitter, Farben, Landschaft, Tierwelt, Pflanzen; Der Mensch
und sein Leben: Korperteile, Familie, Freudenbote, Gesellschaft, Land-
wirtschaft, Kunst und Handwerk', etc.) or of Sabbagh (e.g. 'La nature:
rhomme: les parties du corps humain, les fonctions et l'activite du corps;
la vie sociale', etc.). However, even so primitive a classification of meta-
phorical usage could be helpful, not of course for tracing the literary
education of Muhammad nor for depicting the rustic origins of Islam, but
for semasiological analysis of the scriptural lexicon. 5 Secondly, the same
information might provide a statistical account of formulaic structures and

1 See Wansbrough, 'Note', 55-7, 61. 2 Itqdn, now' 58: iii, 249-89.

3 See above, III pp. 116— 17.

4 See above, I pp. 1-33; III pp. 111-18.

5 See above, pp. 215-16; it is precisely that element which is absent from Allard's
'analyse conceptuelle'.

PRINCIPLES OF EXEGESIS

239

^ systems, and hence a clue to the composition of scripture. 1 Finally, and in
my view of greatest significance, would be an analysis of figure and trope
in terms of archetypal patterns, that is, as the topoi and schemata of mono-
theistic revelation. From the premiss of Bildungserlebnis, in other words,
the material assembled by Sister and Sabbagh, like that made available
in the studies of Horovitz and Speyer, could be profitably pressed into the
service of Quranic form criticism. Bound, as it has been, to the framework
of a very dubious chronology, that same material is unlikely to produce
more than pseudo-history.

An example of archetypal imagery in which, moreover, the source is
quite explicit may be seen in Q. 62: 5 ^} ^ Sljjxll IjL*. ^jJI JJL.
I j liJ J-«^ej j L?JI ^s£~ U> _ji*?t>, as designation not merely of the ignorant

scholar, but also (polemically) of all those unable or unwilling to
perceive the 'true meaning' of God's word: D'HSD NBPU "nan. 2 Now the
Quranic tnathal, is primarily an extended simile, and was classified, some-
what arbitrarily, by SuyutI as either explicit {zahir) or implicit (kamtri). 3
The term itself occurs in scripture eighty-eight times, often with darabna

e.g. Q. 30: 58 (Jl* JS" oT/JI 1^* J trUA) Uj^ JiiJj, occasionally with

sarrafna, e.g. 17: 89 Jl* JS" jl^l lJub J ^U) UTJ^ J^jj. 4 Its basic

function is that of exemplum, and as such mathal may be regarded as
synonymous with aya, hadith, and naba'. 5 That functional equivalence is

stressed in Q. 24: 34 I^Ll. ^jjJI ^ oL--. obi -»5Cj| LJjit OS)j

^JxU! '■ilLc.j^j piCLi y, exhibiting a parallelism of aya and mathal
identical to that of of and mashal in Ezekiel 14: 8 tP'waVl TfWb IrPniBlffni.
On the other hand, the literary character of the Quranic mathal necessi-
tates a distinction between it and the other narrative categories: it is
intentionally anonymous and hence expressly symbolic. 6 Its range is thus
not that of the Biblical mashal, which included taunt, oracle, poem, and
song. 7

It is with acknowledgement by the exegetes of the mathal's symbolic
quality that I am here concerned. Its functional value as exemplum was not
thereby diminished, but rather, and perhaps predictably, enhanced. A point

1 See above, I pp. 47-9.

2 Cf. Geiger, Was hat Mohammed, 90; Hirschfeld, Researches, 94 n. 61; Sister, 'Meta-
phern', 126 n. 2; Speyer, Erzahlungen, 437, 441, 461; Ahrens' proposed parallel with
Matthew 23 : 5, in 'Christliches im Qoran', 165, might almost be described as perverse.

3 Itqdn, now' 66 : iv, 38-45.

4 Cf. Sister, 'Metaphem', 115-16. 5 See above, I pp. 18-20.
6 See Horovitz, Untersuckungen, 7, 25 ; a number of specimens were discussed, always

from the point of view of the prophet's calculated appeal to his audience, by Hirschfeld,
Researches, 83-97; and Buhl, "Vergleichungen', i-ij.

' See Eissfeldt, Einleihmg, 73-100, 106-9; Johnson, 'Mashal', 162-9; the equivalence
hija'lrajaz: mashal as taunt (Spottlied) was noted by Goldziher, Abhandlungen i, 44, 80.

34o QURANIC STUDIES

d'appuiwas provided by Q. 25: 33 (_j*JIj iiUij. Vl Ji»j ttli^jL

l^.t j", in which mathal is antithetically juxtaposed to truth (haqq) but also

to interpretation (tafsir). In the polemical context of that verse, mathal was
traditionally glossed 'falsehood' (butldn), 1 but not without allusion to the
notions of challenge (su'al) and enigma ('ajab). 2 The mathal contained an
invitation to exegesis. From the antithesis mathal: haqq was derived a
number of interpretative procedures designed not only to locate figurative
usage in the text of scripture but also to justify reading there several levels
of symbolic meaning. Such did not ever eliminate entirely haggadic efforts
to connect the mathal with known historical figures (ta'yin), or to identify
the occasion of its utterance (tanzil), of which several not very persuasive
examples may be read in SuyiitL 3 The extent to which exegetes might have
perceived a distinction between historical fact and historical truth, between
Wirklichkeit and Wahrheit, poses something of a problem. For the Tal-
mudic antonyms mashaliemet, Loewe found no evidence of that distinction,
though tnashal itself was one of the thirty-two middot* In Muslim exegesis
a basic 'historical' reference was seldom neglected, though often only as
prelude to excursions into allegorical analysis. 5

As an exegetical instrument mathal might designate rudimentary theo-
logical symbolism derived from imagery so traditional that a consciousness
of figurative usage was not even necessary to its understanding: e.g.
dart' as the unnourishing food of the damned in Q. 88 : 6, or zabad as the
foam or dross of the purifying torrent and fire in Q. 13 : 17. 6 Such was
described by Jurjani as linguistic (lughawt), as opposed to conceptual
('aqlt) coinage: dart remained food, and zabad foam/dross. 7 Ibn Qutayba's
description of both as mathal (the term actually occurs in Q. 13 : 17) may
be thought to have referred not to the words dart' and zabad, neither of
which was metaphorically employed, but to the eschatological context of
both passages. The notion of 'likeness' inherent in mathal rested thus not
upon the apprehension of metaphor, but upon assent to the author's
intention. The 'parable' could be symbolic, even allegorical, but did not
require analysis as metaphor. Related to the technical use of mathal in
exegesis, and the source of some terminological confusion, was the descrip-
tion of certain types of metaphor as tamihil. That practice can be justified
by the semantic element of 'representation' common to most if not all
formations from the root m-th-l, but is none the less misleading. Moreover,

1 e.g. Zamakhshari, Kashshdfiii, 279 ad loc.
1 Cf. Buhl, 'Vergleichungen', u.

3 Itqdn iv, 39-41 ; cf. also Hirschfeld, Researches, 87 n. 8, who could himself not resist
the temptation, e.g. 95 ad Q. 7: 176.

4 Loewe, 'The "plain" meaning', 172-5; see Strack, Introduction, 97 (no. 26); Bacher,
Terminologie i, 121-2, ii, 121.

5 See below, pp. 242-5. 6 Ibn Qutayba, Tdvnl, 49 and 251, respectively.
7 See above, pp. 236-7.

PRINCIPLES OF EXEGESIS

241

the precise nature of the metaphor(s) qualified tamth.il was never satisfac-
torily defined. Zamakhsharl, for example, ad Q. 33 : 72 'We offered (Our)
covenant/trust to the heavens, the earth, and the mountains', sought to
distinguish two kinds of image (tastvir): (a) tamthil, derived from empirical
data (muhaqqaqat), and (b) takhyil, derived from hypothetical data (mafru'
ddt), the two being equally conceivable and equally dependent upon an
exercise of imagination. 1 If Q. 33 : 72 exhibited, in the opinion of Zamakh-
shari, the takhyil variety, other verses admitted of both interpretations,
e.g. Q. 41 : 11 'He addressed Himself to heaven while it was still smoke and
said to it and earth "Come willingly or unwillingly" ', which contained a

trope that could be either tamthil or takhyil: jUlJI ^ jAj

dj^i of (Jst^xil ; 2 or Q. 59: 21 'Had We allowed this

Qur'an to descend upon a mountain you would have seen it humbly
collapse from fear of God', which was both : J^eJj J-^j y*j- 3 It might
well be argued that the operative factor in all three examples is not
metaphor at all, but prosopopoeia/ fictio per sonae* Acceptance of the image
as empirically or as hypothetically derived was not a problem of rhetoric
but of dogma. 5

The role of tamthil as metaphor found better attestation in the tradition
of profane rhetoric. 6 For Jurjani metaphor was of two kinds: (a) those
derived from physical and other sensorily perceptible data whose appre-
hension required no interpretative process (ta'awwuF); (b) those derived
from an intellectual/conceptual (aqli) relation requiring interpretation. He
called the former tashbih, the latter tamthil. 1 An example of the tashbih was
'He is a lion in battle', of the tamthil 'His argument is as clear as the sun',
the clarity of the sun (as opposed to its heat, brightness, etc.) requiring
the additional qualification that nothing come between it and the eye of the
beholder. Description of the Qur'an as 'light' was thus tamthil, and the
word 'light' so employed a mathal for the Qur'an. 8 The basis of tamthilj
mathal was not linguistic and, strictly speaking, not metaphorical, though
confusion may seem inevitable. Use of tamthil as analogy contributed to
that confusion: appearance together of the terms ashbdh and amthdl, as
well as the employment of tashbih and tamthil in the sense of 'assimilation'

J Kashshdf iii, 565.

2 Kashshdf iv, 189.

3 Kashshdf iv, 509.

* See Wansbrough, 'Periphrastic exegesis', 250 (no. 14).

5 See Goldziher, Richtungen, 13 1-4; and cf. Bonebakker, Tatvriya, 24-7 for Zamakh-
shari's use of takhyil.

6 Cf. von Grunebaum, Tenth-century Document, 15 n. 123.

7 Asrdr, paras. 5/1-5, 14/1-3.

8 Similarly, wine might be a tamthil for prophecy, but hardly a 'metapnor', pace
Wieder, Scrolls, 85 n. 3 citing Fasi, Jam? ii, 52.

342

QURANIC STUDIES

(juxtaposition of things similar), are amply attested. 1 That imagery draw
upon 'analogous' formations could not, after all, be thought to represent a
strain upon the resources of any language. But for the terminology of
rhetorical exegesis it is more accurate and convenient to maintain a strict
separation of matkal from metaphor. The latter was bound by linguistic
considerations which could not be, or in practice at least were not,
applied to the range of the former.

In his treatment of mathal Suyuti adduced (anonymously) the following

definition: 2 ol* i'bll J c-jl l^V j*\s>*c& I ej^su JUJt jj^jj Jli^l jU
^-IjptJb I4J (jfcjJI iibcu.^. There the role of the scriptural mathal is ex-
plained as an aid to comprehension, achieved by report to the personifi-
cation of concepts. The reference, in my opinion, can only be to allegory,
of which the prosopopoeic verses adduced by Zamakhshari as tamthilj
takhyil might be held to contain a pale reflection. 3 For those, at least, the
antithesis mathal :haqq is appropriate. The Talmudic application of mashal
also included, in addition to parable, allegorical interpretation, e.g. the
fables of Jotham (Judges 9: 7-20) and Joseph (2 Kings 14: 8-14). 4 Now,
the designedly esoteric characterer of the Quranic mathal was explicit in

the text of scripture (Q. 29: 43): VI lgl2*j L>j ^rUl l$j^*aJ <JL»VI kU)Jj
Oj*JL*JI, an admission of Deutungsbedurftigkeit comparable to Matthew
13: 10-13.5 The assumption of those exegetes not concerned with identi-
fication of dramatis personae or with relation of the mathal to a remembered
historical event (real or Active) was of emblematic language, by means of
which levels of significance could be discerned in scripture.

These levels were not mutually exclusive, but rather, parallel and com-
plementary. Ultimately incorporated into standard works in the exegetical
tradition, that principle was concisely set out in the introduction to the

Tafsir of Sahl Tustari (d. 283/896) : 6 hoj l^Jj VI jT,2JI ^ JuT ^ Lj

«Ull y Lj2» Ljj i\yJ\ (Je «_J^Jt <3\j£\ ^ikJIj Every Quranic.

verse had thus four 'meanings' : zdhir (literal), batin (symbolic), hadd (pre-
scriptive), and matla' (spiritual). My translations are only approximate: in
view both of their number and order of appearance correlation with the
quadrivium of medieval Biblical exegesis may be justified: 7

1 See above, pp. 166-7; Goldziher, ZSKriten, 104-5; TahanawS, Iftildhdt, 1193-4.

2 Itqdn iv, 38. 3 See above, p. 241. 4 Bacher, Terminologie i, 122.

5 Also Mark 4: 10-3, Luke 8: 9-10; one of these passages, probably Matthew 13, was
mentioned by SuyutI, Itqdn iv, 39.

6 GAS i, 647 ; Tafsir, 3 ; adduced anonymously and abbreviated in Suyufl, Itqdn iv,
196-7; see above, III pp. 104-5.

7 See Lausberg, Handbuch, para. 900 (according to Rabanus Maurus); Richter, Exe-
gese, 15, 174-90-

PRINCIPLES OF EXEGESIS

243

zdkir: historia
batin:allegoria
kadditropologia
maila: anagoge

While, in the Muslim tradition, the rich possibilities of polysemy and multi-
valence had already begun to be exploited by elaboration of methods
derived from the principles of wujuh and tawil, 1 it is worth noting that
the schematic arrangement of four levels of 'meaning' for every ( !) scrip-
tural verse exhibited considerable refinement of the earlier binary opposi-
tion muhkam:mutashdbihjmusktabih, in which theoretically (at least), only
the latter were susceptible of more than one 'correct' interpretation. The
fourfold system was, moreover, first formulated and invariably advocated
by exegetes whose concern with the literal sense (historia: secundum
litteram), even when expressly declared, was minimal. Here devoid (or
nearly so) of the polemically charged connotations of halakhic usage, 2 the
term zahir was reduced to little more than a point of departure for symbolic
and eschatological speculation. Coexistence of four semantic values implied
both equality and independence of function: Jerusalem was thus the
capital of the Jews {historia), the church of Christ (allegoria), the soul of
man (tropologia), and the city of God (anagoge).* The Muslim designation
of that phenomenon was tatbiq, described by Suyuti as a kind of symbolic

parallelism: 4 jj'U^ ,J| ILii. oljUl l$J iiUi £>j Uyfcljt J* ^j^aJI. An

example is the interpretation imposed upon Sural al-Fil by the Sufi Ibn
'Arab! (d. 638/1240): in the historical attack of Abraha on the Meccan
sanctuary was reflected the assault of the powers of darkness upon the soul
of man, and in its repulsion deliverance of the soul from the snares of
fantasy by the powers of intellect* The exegetical principle itself might be
described as tropologia, and its relation to historia defined as tatbiq.*'

Symbolic parallelism is the necessary substratum of all allegory, as well
as of irony and caricature. Its success required uninterrupted consciousness
of the literal 'ground', the source of whatever persuasive power the imagery
of superimposed levels (whether allegoria, tropologia, or anagoge) might
possess. The relation is one of counterpoint, present in allegory both as
creative mode and as exegetical device. 7 It might not be unjustified to see
in Philonic allegorism an identical set of postulates, for which paremphasis
expressed the counterpoint between literal (phaneros) and symbolic

1 See above, pp. 154-6, 208-12. 2 See above, pp. 150-1, 152-3, 187-8.

3 Lausberg, Handbuch, loc. cit.

4 Itqdn iv, 195; read possibly haqaiq for daqd'iq.

5 Cited Goldziher, Richtungen, 242-4; see above, I pp. 42—3.

6 Pace Goldziher, loc. cit., who contrasted tatbiq with ta'wil (the latter described as
'wirkliche Allegorie'), following Ibn 'Arab!.

7 Cf. Frye, Anatomy of Criticism, 89-92 : the 'contrapuntal technique'.

244

QURANIC STUDIES

(hyponoia) levels of significance. 1 In Suyuti's definition of tatbiq, the
elements linking literal expression (nususjzawakir) with arcane meaning
(daqd'iq : Ihaqd'iq) were designated 'concealed allusions' (ishdrdt hhafiyya),
or better 'signs'. 2 The linkage which followed upon apprehension of the
sign was called i'tibar (transition). 3 Unlike tatbiq and ishara, which were
corollaries of the agreement to recognize manifold 'meaning', i'tibar was a
reference to procedure. 4 It was not, however, i'tibar, but ta'wil which
became the generic designation of symbolic exegesis. Reason for termino-
logical development lay, of course, in the close association of ta'wil with
the concepts of polysemy/homonym (wujah). The antithesis ta'wil: tafsir
acquired new significance. From an almost neutral description of rational,
as contrasted with traditional, interpretation, 5 ta'wil became first a collec-
tive expression for all save literal exegesis (zahir), and finally an epithet of
abuse for irresponsible, as contrasted with 'respectable' scriptural exegesis
(tafsir). Polemical reference to ta'wil was nearly always abusive, e.g. in
the writings of Ibn Hazm. 6 Among practitioners of ta'wil, the term tafsir
described the necessary first step (historia) of any interpretation, but no
more than that, e.g. in the work of Ibn 'Arab!. 7 Patronization of tafsir by
Sufi exegetes found a complement in the criticism of their methods by
opponents who perversely rejected Sufi exegesis precisely because it was
not 'tafsir'*

In Tustari's work symbolic interpretation is primitive and archetypal.
The equivalence scripture: light (qur'dn.nur) was, for example, justified
by reference to the intermediate term 'guidance' {kudo), derived explicitly
from Q. 42 : 52 and implicitly from 24 : 40. 9 Ad Q. 2 : 269 wisdom (hikma)
represented self-discipline in adversity, elimination of carnal appetites,
and spiritual vigilance. Wisdom also comprehended all of the sciences, the
basis of which was Sunna. Similarly, knowledge ('Urn) was essentially
arcane, and those granted access to it (al-rdsikhun fil-'ilm) the special reci-
pients of divine favour. 10 Ad Q. 24: 35 the image illumination/wisdom was,
not unexpectedly, elaborated in some detail: i.e. divine light, the light of
Muhammad, the heart of the believer incandescent with the illumination

1 See Loewe, 'The "plain" meaning', 143-51, esp. 148.

2 Goldziher, Richtungen, 225-7: 'Hindeutungen'.

3 Goldziher, op. cit., 245-51 : 'Hiniiberschreiten'.

4 Pace Goldziher, loc. cit., where i'tibar is defined as the halakhic application of tatbiq.
s See above, pp. 154-5.

6 Goldziher, Zdhiriten, 132 n. 2; id., Vorlesungen, 108, 159.
' Goldziher, Richtungen, 224-57, esp. 239 n. 2.

8 See the discussion in SuyutI, Itqdn iv, 194-8; that view was to some extent shared
even by moderate Sufi exegetes, e.g. Suhrawardi, see Goldziher, op. cit., 186 n. 1;
Jullandri's unqualified description of Sufi exegesis as 'tafsir' is simplistic and misleading,
as is his distinction between 'symbolic tafsir' {ishdrijramzi) and 'speculative tafsir' (tiasfori),
cf. his study 'Qur'anic exegesis', 105-19.

» Tustari, Tafsir, 5.

10 Tustari, 32-3, and 36-7 ad Q. 3 : 7.

PRINCIPLES OF EXEGESIS

245

of divine unity, and finally, the Qur'an as lamp: knowledge its light, com-
mandments its wick, and purity its oil. 1

It is precisely such straightforward substitution/transfer which character-
ized this earliest symbolic exegesis: e.g. umm al-qura in Q. 42: 7 was both
Mecca (zdhiruka) and the human heart {batinuha), its environs the members
of the human body; lawh mahfUz in Q. 85 : 22 was the breast of the believer,
in which truth might abide. 2 Underlying interpretation of that kind was
the acceptance of extended simile : the extent to which it may justifiably
be described as allegory depends upon the nature of the scriptural passage
subjected to exegesis. Tustari's work contains almost no commentary on
the 'narrative' sections of the Qur'an, that is, those which would lend
themselves most easily to the action: imagery transfer typical of allegory
(e.g. Suras 12 and 18). The technique of dramatic allegorization found later
and full expression in the writings of Ibn 'Arab!. 3 Earlier traces may be
seen in commentaries ascribed to Ja'far al-Sadiq (d. 148/765) where, for
example, the ascension of Muhammad (mi'raj al-nabi) was analysed as a
threefold passage: from Mecca to heaven (malakut), from Medina to power
(jabarui), and from birth back to his creator. 4

Recourse to symbolic interpretation was very much a characteristic of
sectarian exegesis, for an important part of which the names of Ja'far and
his father Muhammad al-Baqir were significantly authoritative. But a more
appropriate, because datable and indisputably authentic, illustration of that
particular technique is found in the Tafsir of Qummi (d. 309/921). 5 There,
ad Q. 14: 24-6, the mathal contrasting the good and the evil word (kalima)
with the good and the evil tree (shajara) was interpreted as reference to the
contrasting histories of the prophetical progeny (ahl al-bayt) and the dis-
believers (kdfirUn), with concomitant elaboration of the imagery derived
from root, branch, and leaf. 6 Ad Q. 15 : 87 the seven matham were under-
stood to refer to the Shi'I imams; 7 and ad Q. 18: 60-82 the long dialogue
between Khidr and Musa was related to a prognosis of Muhammad's
appearance as herald of the true faith. 8 The device by which agency in the
narratio was transformed into imagery appropriate to the Islamic theodicy
could be construed as allegory, but because of the specifically historical
mention in such exegesis, it may more accurately be described as typology. 9
The historicization or actualization of scriptural imagery is the converse
and complement of allegorical interpretation, and both require assent to

1 Tustari, 103. 2 Tustari, 128-9 and 180, respectively.

3 See Goldziher, Richtungen, 233.

4 GAS i, 528-31, e.g. 529 no. 2; MS Nafiz Pasha 65, 7o v -i r ad Q. 17: 1; see above,
II pp. 67-9.

5 See above, pp. 146-7; and Goldziher, op. cit. 279-309.

6 Qummi, Tafsir i, 369.

7 Qummi, i, 377; cf the emendation/exegesis ummaxa'imma, above, p. 167 n. 4.

8 Qummi, ii, 37-40; see above, pp. 127-8.

* Cf. Lausberg, Handbuch, para. 901; Seeligmann, 'Midraschexegese', 167-76.

QURANIC STUDIES

the symbolic quality of the schemata of revelation. Definition of those
schemata as the projection of cultural and spiritual ideals into history, or
as the refraction of history in poetic imagery, will depend upon the use to
which they are being put: whether as rhetorical device (synthetic) or as
exegetical technique (analytic).

In Biblical literature the terms employed for typological exegesis
originally designated the interpretation of dreams, i.e. patar and peshar. 1
The 'prognostic exegesis' of sectarian Judaism, Qumranic and Karaite,
consisted exclusively of typological equivalents drawn from different but
allegedly parallel historical processes. 2 While it is certainly tempting to see
in the Islamic term tafsir a reflex of patar/peshar,* the literary evidence
provides little support for the conjecture. The hapax legomenon in Q. 25 :
33 referred not to dream nor to scripture, but to mathal, and the origin of
tafsir as terminus technicus belonged to the tradition of profane rhetoric. 4
Now, the Quranic equivalent of p-t-r in Genesis 40: 8 and of p-sh-r in
Daniel 5 : 12 is tawil, which occurs eight times in Sural Yusuf, always
glossed 'dream-interpretation' (ta'bir al-ru'ya), and nine times elsewhere,
glossed 'outcome'/'sequel' ('dqiba)* The eschatological and prognostic
overtones of both uses render ta'ml an appropriate designation of typo-
logical exegesis. That the Muslim term tafsir might, on the other hand,
have reflected a characteristically sectarian and polemical emphasis upon
the recent fulfilment of a historical promise articulated in Hebrew scrip-
ture remains a distinct possibility. Such, indeed, was the function of all
scriptural interpretation and the task imposed upon exegetes of every
allegiance :

137 i»b »Tr> m aorm 'a

(Qohelet8: 1)

1 Bacher, Terminologie ii, 177-80, 173-4, respectively; Gertner, 'Terms', 17-18.

2 See Wieder, Scrolls, 199-213; Rabin, Qumran, 117.

3 Rabin, Qumran, 117. 4 See above, pp. 154, 233-5.
5 See above, pp. 156-7.

s

INDEX OF NAMES AND SUBJECTS

'Abdallah b. 'Abbas, 36, 69, 95, 120, 133,

HO, 143, 146, 150. 155, 158, 163-5, 179.

185, 192, 203, 216, 219
Abraham, 4, 8, 19, 20, 21, 37, 49, 54, 56, 77
Abraham b. David, 201
Abrogation, 180, 192-202
Abu 'Ubayd, 50, 150, 155, 188, 193-4,

198-9, 203-4, 21 9
Abu 'Ubayda, 75, 83, 219-21, 223, 22 5>

228-9, 235, 237
Ahl al-kalam, 174
Analogy, 148-70, 175, 190, 208
Angelology, 30-1, 58-63
Ash ari, 81-2
'Ashura', 181, 183

Baghdad!, 199, 201

B. 'Urayna, 185-8

Baqillam, 80, 84, 93, 232

Beduin, 94—6

Bukhari, 60, 181, 183, 217

Canonization, 43-51, 77-84
Companions, 203-25
Companion codices, 203-7
Contradiction, 161-6
Covenant, 8—12, 16

Dani, 169, 204-5
Dinawari, 146

Distributional analysis, 212-16

Eloquence, 93-9, 162-3, 221

Ethnocentricity, 53-4

Evangelium, Muhammadan, 56-73, 76, 83,

176, 194-5
Exile, 7-8

Farra', 25, 45, 94, 120, 133, 143, 155, 200-7,

218, 228
Figurative language, 227—46
Formulae :

apodictic, 13-16

commission/messenger, 12-13, 24, 53
narrative, 18-20
supplicatory, 16-18
Formulaic language, 48, 1 17-18

Gabriel, 12, 30-1, 34, 46, 61-2, 63-5, 70,

122-3, I2 6, 181
Gospel, 37, 63, 72

Hasan Basri, 160-3, 2 °5
Hibatallah, 184, 196-7
Hisham b. Kalbi, 16-17
Holy War, 170-3, 183-5
Hudhayl b. Habib, 144, 172, 174

Ibn Abl *1-Isba\ 232-3, 237

Ibn Anbari, 101, 168, 221

Ibn 'Arabi, 151, 185, 189-90, 218

Ibn Hisham, 9, 122, 217

Ibn Ishaq, 57-8, 122-7, l8 3

Ibn Kammuna (Sa'd b. Mansur), 200

Ibn Masud, 203, 207, 222

Ibn MuqafTa', 159-60, 166

Ibn Mu'tazz, 232

Ibn Qutayba, 74, 107, 146, 151, 165-6, 219,

222-4, 228-31, 240
Inflexion, 104-5, 106-11
Inimitability, 77-83

Ja'far b. Abi Talib, 38-43, 47
Ja'far al-Sadiq, 146, 245
Jahiz, 73-4

Jassas, 151, 156, 183, 188-9, 191-2. 196,
217

Jesus, 8, 37, 55, 63, 65, 66, 71, 74, 195
Jurjani, 80, 229-30, 236-7, 240-1

Ka'b Al-Afcbar, 75, 189

Kalbi, 34, 59. 62, 115, 130-7, i4°-6, 149-

50, 183, 206
Kisa'I, 45, 212-15
Koine, 87-90

Lexicography, 216—19

Lingua sacra, 99-118, 208, 210, 215, 231

Litany, 26

Logia, 1, 4°, 4 2 , 47, 49-5 2 , 78, 148
Logos, 77, 1 20-1

Ma'arri, 94, 96

Maimonides, 64, 74, 187-8, 201

Malawi, 165, 211-12

Malik b. Anas, 58, 171-2

Maturidi, isi, 150, 154, '56, 177

Moses, 4, 8, 10, ii, 19, 20, 29, 35-7, 49,

55^7, 65-6, 68, 71-4, 76, 78, 83, 99, 195
Muqatilb. Sulayman, 3, 69, 122-36, 140-6,

149-50, 163-5, 170-4, 183, 206, 208-11,

212, 218
Muslim, 180-1, 183, 217
Mu'tazila, 79, 81-2, 84, 160, 215, 236

348

INDEX OF NAMES AND SUBJECTS

Nahhas, 184, 193, 196-8
Narrative style/structure, 127-48

Papyri, 90-2

Pericopes, 2-5, 12, 20, 42, 49, 53
Periodization, 48, 1 16-17
Poetry, Jahili, 94-8, 102, 105, 142, 216-19,
226

Polysemy, 208-12

Prescription, 172-202

Prophethood, test of, 122-6, 146, 195, 198

Pseudo-'Afa Khurasani, 186

Pseudo-correction, 87-8

Qartajannl, no
Qasim b. Ibrahim, 160
Qazwini, 232-3
Qummi, 41, 146, 245

Remnant, 4, 11

Restoration, 219-26

Resurrection, 31-3

Revelation, modes of, 34-8, 58-61

Saadya, 5, 9, 35, 37, 77, 84, 200

Sanctuary, 16-17, 42, 69

Schemata, 1-33, 47, 100

Sectarian milieu, see Canonization

Shafi'i, 151, 168, 174-5, 193. 196

Sharif Murtada, 217, 227, 235

Shu'ayb, 21-5, 28-9, 42, 235

Sibawayh, no

Sign, 5-7, 18-19

Sira, 39-42. 57-8, 121-9

Stoning verse, 70, 193-6, 198

Sufyan Thawri, 137-8, 140-5, 183, 206,

212, 2l8

Sulami, 180, 198-9

Sunna, 51, 56^-7, 78, 159, 161, 163, 172,
174-6, 188-90, 193, 198-9, 202

Suyflti, 14, 35, 41, 43, 46, 51, 55, 59-60,
93-4, 104-6, 108-10, 116, 126, 136, 152,
165, 169, 176, 178, 180, 185, 204, 205,

211, 2l6, 2l8, 219, 221-2, 224, 237-9,

242, 244

Sword verse, 184-5, 197-8

Tabari, 59, 73, 139-40, 146, 155, 186, 197,

217, 219
Tha'lab, 143
Thaumaturgy, 73-83
Theodicy, see Umam khaliya
Tirmidhi, 182-3, 217
Torah, 36-7, 64, 75, 79, 83, 195
Tradition, 179-83
Tustarl, 104-6, 242, 244-5

Ubayy b. Ka'b, 203, 207, 222
Umam khaliya, 3-4, 10-n, 21, 29, 53, 124,
160

'Umarb. Khattab, 57, 158-9, 179, 195, 217
Uncreated Qur'an, 77-84
'Uthman/'Uthmanic, 43-6, 49, 51, 115,
158, 203-4, 221

Variant readings, 202-7
Variant traditions, 20-8, 207

Wahidi, 178-9
Zajjaj, 150, 155

Zamakhshari, 12, 26-7, 35-6, 59, 62, 98,
101, 108, 113-14, 115, 120, 139, 152,
154, 169, 206, 224, 242

Zuhri, 180, 198-9

INDEX OF TECHNICAL TERMS

Shad, 204
'ahd, 9-12, 16
ahruf, 45, 73, 204

akhbar al-ghayb, 69-70, 79-80, 144-5
alqa, 60 , 196
amana, 10

'amm/'umum, 169, 178, 185, 190
amr, 77

apostrophe/parenthesis, 28

'arabi/'ajami, 94, 98-9

asbab/akhbar al-nuzul, 38, 41, 123, 127,

132, 141-2, 177-8, 179-83, 196
a§l/usul, 158, 162, 170
aya, 6, 18-19, IS3

ayyam (allah/al-'arab), 4, 29, 125, 140

bada', 197, 199, 201
badi', 234, 238
batfn, 15 1-3
bayan/bayyina, 153
berit, 9, 11

circular explication, 130-1, 145
corroboratio, 13, 24

dabar, 77

dala'il al-nubuwwa, 70, 73
Deutungsbediirftigkeit, 100, 118, 148-54,
242

dhimma/dhimmi, 10, 194
diraya, 154, 227

ellipsis, see hadhf
erlebte Rede, H4

exemplum, 4, 18-21, 24, 47, 51-2, 58

fasaha, 80, 93, 97, 201
fasila, 116
fatra, 54

gazerah shawah, 167
gloss, see interpolation

habl, 10

hadd/hudud, 177, 195

hadd/mafla', 242-4

hadhf, 181, 227, 229

hadith/h. qudsi, 129, 133, 145, 179

halal/haram, 173, 191

haqiqa-majaz, 236-7

hebin, 153
hijab/satir, 34-5, 73
hiqqish, 167

liistoria/allegoria/tropologia/anagoge, 242-4
hukm/ahkam, 176-7, 183-8, 195, 201
hyperbaton, see taqdim

ibtida', 152, 224
identification, see ta'yin
idgham, 181
!jaz, 227

i jaz al-qur'an, 79-83, 231-2
ikhtilaf, 163-6, 183
ikhtisar, 227, 229
ikhtiyar, 223, 226
ilhan 1 . 34i 58-60, 210
'iUa, 168

imam, 54, 184, 187

interpolation/paraphrase, 27

i'rab, 95, 104-5, 106-11, 155, 221-2, 229

irsal, 34, 58-60, 210

irtibat, 215

ishara, 244

'isma, 67, 71

isnad, 179-83

isra'/mi'raj, 66-9

isti'ara, 229

istilah, 103-4

isti'naf, 143, 152, 224

i'tibar, 244

jam', 46

jarh wa-ta'dil, 140
jihad/qital, 170-3
jumla, 36

kalima/kalam (allah), 76-7, 80, 103-4,

120— 1
kelal/perat, 169

khass/khusus, 169, 178, 185, 190
khatam, 64-5
khutba/khatlb, 147-8
kinaya, 231

kitab/kutub (allah), 57, 74-6, 161
kitba/lafz, 223
kitman, 63, 71, 189
ktib/qere, 223

laff wa-nashr, 233-5, 238
lafz/malfuz, 189, 226

INDEX OF TECHNICAL TERMS

lahn, in, 221, 228

luhun wa-aswat, 104-5, rl1 . 221

madhhab kalami, 232-3, 238

majaz, 115, 120, 169, 219-26, 227-38

mashal, 239-40

mashhur, 206

mathal, 239-45

matluw/ghayr m., 176

metaphor, 238-42

minimal units, 130-1

mithaq, 8

mithla/ka (qawl), 127, 142-3, 167, 212, 223

mu'arada, 81, 160

mubalagha, 230

mubham, 136

mufassal, 234

muharibun, 185-7, '93

muhkam, 149-S7. 163. i7°> I73> W, 184,

190, «43
mujmal, 234

munajjam/nujum, 36-8, 230
munasaba, 215
musaddiq, 65

mushaf/masahif, 43-5, 51, 203-4
mushtabih, 142, 157, 165, 208-9, 212-15,

220, 243
mutabaqa, 147

mutashabih, 142, 149-57, 163, 190, 192,
208-9, 212, 243

narratio, 42-3, 47, 127-47, 171. *77> 226,
245

naskh, 47, 150, 183, 191, 192-202
nass/mansus, 168, 187
nazir/naza'ir, 127, 142-3, 169-70, 211-12
naztn, 79-80

panim, 155

paremphasis, see tatblq
patar, 246
peresh, 154

periphrasis, 220, 225, see majaz
peshar/pesher, 50, 246
peshat/peshut, 152
prolepsis, 128

qafiya, 116
qarina, 116
qa?5, 147
qira'a, 204

qiyas, 159, 166, 168, 190

rajih/marjuh, 155, 187
rasm/khatt, 204
ribat, 185
riwaya, 154-6, 227

ruh, 122-6, 181, 235

sahlfa/suhuf, 9, 78
saj*, 1 1 6, 147
sarfa, 79-82

serial repetition, 130, 145

shabah/ashbah, 142, 167, 169-70, 190, 209,

241
shadhdh, 206
shaytan, 59-61
?ifat, 83

simile, see tashbih
stage-direction, 129, 145
subnexio, 234, see tafsir
supercommentary, 129, 203

ta'awwul, 236
tab', 64-5
tabaqat, 140, 158
tabdil, 189, 196
tadadd, 165, 228
tafanun, 215
tafstt, 75, 153-4

tafsir, 57, 121, 127, 154-8, 191-2, 234-5,

244, 246
tahaddi, 7, 79-82
tahiannuth, 18
tahlil, 17-18
tahrif, 63, 162, 189-90
takalluf, 80
takbir, 18
takshif, 154

takhsis, 188, 191-2, 201
takhyil, 241-2
takrar, 178, 230
ta'lif, 79-80
talbiya, 16-17
ta'lil, 168, 191
ta'lil takhyill, 237
tamthil, 240-2
tanaqud, 163, 165, 228
tanwin, no

tanzil, 59, 84; see asbab al-nuzul
taqdim, 143

taqdir, 4, 115, 169, 219-26, 227-8
tarannum, m-12
ta'rib, 219
ta'rid, 231

tarjih, 187, 206, 226
tartlb, 46, 80
tartll, 36
tasbih, 17
tasfir, 154
tashbih, 241
tasmiya, 16, 18
tasrih, 231
taswir, 241

INDEX OF TECHNICAL TERMS

tatblq, 243

tawafuq/tawarud, 219

tawatur/mutawatir, 204-5

ta'wil, I2i, 152, 154-7, *%9> 244, 246

tawqif, 103-4

tawriyya, 231

ta'yin, 30, 135-6, 141, 145. 164
tilawa, 176, 196

umm al-kitab, 149, 153, 170
ummi, 53-4, 63

Urtext, 43, 46

varia lectio, 21, 44~7. mi, i32> 138, 144,

202, 206, 221
vaticinatio, 69, 145, 176

wahy, 34, 58-60, 209-10

wajh/wujuh, 142, 151, 155, 208-11, 235

z5hir/azhar, 151-2, 187, 242-4
ziyada, 229, 236

INDEX OF QURANIC REFERENCES
(SELECTIVE)

I

153

2: 206

13

I

4

14, 108, 205

2. 207

13

2

213

3: 3

/

3: 7

149-53) 156, 17°, 224,

2

25

235. 244

2*7

3: 13

6, 19

2

38

o A

3: 20

14, i7

2

3°

18

3= 3°

54

2

3 1

1 01
iu j

3: 40

4

214

3: 45

30

*fr.S

AO

3- 64

40

2

49

4

3= 67

54

2

59

109

3: 79

75

2

DO

25

3: 81

8

2

„o

78

54

3: "2

10

2

02

27

3: 117

"3

2

Q_
*»7

13

3= 137

8

2

00

72

3: 141

207

2

97

3: 154

39

2

102

3°

3= 155

170

2

I Ou

196, 201

3: 161-3

170

2

III

234

3: 169-70

170

2

117

77

3: 188

17

2

I20

*4

3: 190

5

2

124

IOO

3: 194

17

2

J 34

2

3: 200

170-2, 185

2

I43 _ 4

142

2

154

T*7A

4* 10

40

2

181, 182

4: 13-14

177

2

I04

32

4: 15

40

2

173

15, 4O

4= 23

164

2

l83

40

4: 36

40

2

1 04

IOI

4: 44

19

a

1 8s

37. 234

4: 46

76

2

196

40

4 - 47

13

2

216

170-2, 184

4: 74

170

2

2 i 8— 22

7. 14

4: 93

180

2

229-30

177

4: 95-6

170

2

232

181

4: 100

7

2

234

181

4= 125

55

2

238

204

4: 153

74

2

248

6, 19

4: 163

209-10

2

252

20

4: 164

35. 223

2

253

35

4: 171-2

30. 4i. 55. 60, 77

2

259

32

2

265-6

27

5: 3

40

2

282

25

5=4

2

285

55, 76

5= 7

9

INDEX OF QURANIC REFERENCES

5: 13

72, 189

5-' 14

1 1

5: 27

20

5: 33

185-8, 191, 228

5: 43-Si

192-5

5: 50

39

5: 96

32

5: in

209

5: 153

17

0:

2

0: 15

14

6 \ 19

35. 209-10

6r 22—3

104

25

2

A*

o . 4"

19

0: 55

13

C. . O ^
O: 02

169

6: 95

32

O: IO4

14

0: 109

115

0: 112

6l, 209-IO

0: 114

14

6: 115

76

6: 118

71

6: 121

l8, 71, 200-I0

6: 131

2

6: 145

40

6: 152-3

25, 40, 177

6: 154

175

6: 161

114

7: 2

"5

7: 4

2

7: 19

214

7: 26

15, 237

7: 33

40

7= 34

2

7: 43

159

7: 46

35

7: 69

20

7: 70

40

7: 85-93

2i-5, 235

7: IOI

19

7: 103

19

7: 114

30

7= 143-5

35-6

7: 157

54, 63

7: 158

15, 120

7: 169

9

7: 185

19

7: 187

H

7: 188

65

7: 203

7

8: 5

8

8: 12

60

8: 15-16

170-2

8:38 2

8: 41 170-2

8: 56 9

8: 60 171-2

8: 65-6 170-2, 196-7

8:72 9

9: 1

11

9: 3

I OO

9: 5

I84-5, I9I-2, 197-8

9:6

83

9: 8-10

IO

9: 25

170-2

9: 29

170-2

9: 30

123, 231

9:38

8

9: 70

19, 25

9: 72

27

9=75

9

9: 97

177

9: in

170-2

9: 112

177

10: 9

27

10: 13

2, 32

10: 20

7

IO: 33

13

I0: 34

31

10:47

2, 53

IO: 67

6

10: 102

4

IK 29

39

Ii: 42

15

11: 49

19

11: 62

40

11: 81

8

11: 84-95

21-5

11: 102

13

11 : 116

4

12

130-9, 157. 245

12: 2

62

12: 3

19

12: 17

114

12: 47-8

114

12: 59

25

12: 66

9, 28

12: 67

15

12: 74-5

114

12: 82

220, 225, 227, 228-9

12: 99

145

12: in

3

13: 5

3i

13: 6

4

13:8

53

254

INDEX OF QURANIC REFERENCES

14

18: 73

231

13: 37

99

18: 79

203

13: 39

153. 201

10: 03

14

13: 4i

19

18: 107

143

14: 4

53. 98, 99, 218-19

19:7

35

14: 5

4. 19. 29

19: 10

207

14: 35

39

19: II

209—10

14: 41

33

19: 16

'9

19: 17-21

12—13

15: 4

a

1 9 • 4** "

7

15: 10

2

Tft* eft

An
4 V

15: 54

108

19: 59

162

15: 87

2, 245

19: 64

r 4

15: 9«-3

164

19. 70

9

19: 97

99

16: 6

6

16: 8

171-2

20: 9

19

16: si

28

20: 27

99

16: 63

2

20: 39

OO

16: 68

209-10

20: 55

31

16: 91

10

20: 63

108, 222-3

16: 97

164

20: 99

2

16: 103

98

20: 114

37

10: 113

53

•7ft* TIC

1 1 — 12

16: 115

40

20: 123—4

IO4

16: 120

54

21: 5

7

17: 1

8, 67-9, 99, 120

232—3

17: 2

36

21: 35

4

17: 22-39

177

21 : 107

54

17: 34

9

17: 35

H3

22: 30

40, 72

17: 45

35. 72-3

22: 39-40

170—2

17:48

19

22: 52

60, 198, 201

17: 55

55

23: 8

17: 60

68

10

17=85

124-6, 181

23: 14

32

17: 86

34

23: SI

2

17: 88

79

23 : 02

32

17: 89

239

23 . OS

40

17: 90-3

7,68

23: 101

104

17: IOO

30

23: 115

ID

18

122-9, 245

24: 2

193

18: 6

217

24: 4

40

18:9

3,6

24:23

40

18: 14

217

24: 34

24O

18: 16

7

24: 35

41

18: 19

13

36, 120

18: 22

143

25: 32

18:31

142

25: 33

154, 240, 24<

18: 32

20, 27

25: 46

28

18: 38

181

18: 48

31

26: 42

3°

i8:55

3

26: 176-90

21-5

18: 69

143

26: 194-5

99

INDEX OF QURANIC REFERENCES

26: 210

59

26: 221-7

60

27: 1

114

27: 34

3» 12-13

28: 34

99

28: 49

83

28: 73

234

28: 75

53

29: 3

4

29: 4

16

29: 17

40

29: 27

75

29: 36-7

21-5

29: 43

242

29: 45

18

29: S3

15

29: 56

17

30: 1-4

7°. 144. 176

30: 9

19

30: 12

33

30: 23

234

30: 40

32

30: 42

3

30: 43

4

30: 47

13

30: 56

32

30: 58

239

31: 12-13

16

3i: 13

169

31: 28

113

32: 14

4

33: 7

8, 11

33= IS

9. »

33= 21

175

33-- 33

39

33= 4°

64

33 : S3

35

33= 63

28

34: 15-16

27

34: 19

4i 19

34: 3°

4

34= 50

14

35= 3

15

35= 24

53

35:28

108

36: 76

107

36:78

32

36: 79

31

37: 27

164

37: 35

T ft
I O

37: 36

60

37:79

20

37:80

13

37: 157

75

IT lf\A~-fi
J 1 • 1 U, T u

J 4

37: 168

7

38: 13

25

30. 17

*9

38: 29

6

38:6s

14

39: 28

99

39: 64

108

39:68

33

40: 15

60

40: 78

7

4i: 5

35

41: 9-11

164

41 • 12

34, 209-10

41: 17

108

41: 43

55

41: 44

_Q
98

42: 17

42: 51-2

34-6, 58

43: 4

153

43- 8

2

44: 23

Q
O

45: 14

4

46: 9

55

47: 22

40

48: IO

10

48: 15

76

48: 26

39

48: 29

6

49: 9-IO

170-2

50: 14

25

SO! 15

32

50: 29

76

50: 42

32

5i: 15

27

51: 3°

13

51: 5°

8

256

INDEX OF QURANIC REFERENCES

Si: 52

60

79: 27-30

164

ki: IO— 22

JO* f 7 ****

108

80

107

53 : 55 <

26

80: 3

28

54.: 17

99

82: 17-18

28

54:43

11

83: 2-3

25

33 • *j

6

83: 6

33

K-Xt 27

2^7

83: 19

28

33 *

164

83:21

3°

C e * a6 — 7*7
33 • / /

2C— 7

83:28

3°

s6: II

30

85: 21-2

83, i7«>,

56: 62

32

56: 75

36

86: 8-9

32

56: 88

30

89: 15

4

58: 4

177

90: 12

28

8

91 : 8

lA. CO

61: 4

170-2

61: 6

17,64

92: 17

1-78
I70

61: 10-13

170-2

94: i-3

66-7

62: 2

39

62: 5

2*10

96: 18

30

6?: 1

177
11

97: 1

37, 62

68: 17-33

27

99: 5

209-10

69: 12

■ 7

171 — 2

19

691 24

A

101 : 8-11

27

69: 41-2

00

73: 10

7

105

42, 225

74: 27

28

106: 1

225

74- 3 1

4

112

14, 18

75: 5

164

75: 19

153

"3

14

77: 14

28

114

14

ANNOTATIONS

THESE annotations have been included in this reprint in order to
remove some of the barriers to understanding Quranic Studies for
those whose Arabic, Judaeo-Arabic, Latin, Hebrew, German, and
French is limited. A good deal of acquaintance with Arabic remains an
absolute necessity to make sense of this book, certainly, but through a
combination of transliteration and translation, it is hoped that the
reader will find the material somewhat more accessible. Where the
point of the Arabic script citation in Quranic Studies is a linguistic one,
transliteration is provided; translation is provided where the point re-
lates more to content. It must be admitted, however, that the dividing
line between these two is not always clear; thus, on occasion, clarifica-
tion is provided by transliterating individual words in the context of
translation into English. Arabic words that already appear in trans-
literation in the text have generally not been translated. Passages in
Arabic script that are already translated in the book have been noted as
such. Published translations have been used where possible but have
often been modified to accord with Wansbrough's intention behind
adducing the passage.

The glossary provides meanings for instances of French, German,
Greek, and Latin terminology that are not glossed in the text or that
are used on multiple occasions, as well as for English words that caused
me difficulty when I first read the book. The latter category is highly
subjective, and I recognize that everyone will have a different sense of
what should be included or what could (even should) be excluded.

Bibliographical references for Arabic manuscript works that have
been published since the publication of Quranic Studies have been pro-
vided in the new section that follows the bibliography to this book. At
one time I considered providing page references to the published
editions in these annotations; however, in the end, I decided that those
readers who have the necessary skills to consult the published works
will be able to find the corresponding page numbers reasonably quickly
on their own initiative.

Words in brackets indicate a correct reading and that there is a
misprint in Quranic Studies; words in parentheses generally indicate
matter added for clarification. In passages translated from exegetes

258 ANNOTATIONS

(etc.), quotations from the Qur'an have been underlined in order to
clarify the separation between text and gloss, although it is important
to remember that such clarifications are not always apparent in the
original text.

These annotations and glossary have been placed at the end of
Quranic Studies in order to maintain the integrity of the book which
John Wansbrough wrote. The translations provided here may not
always convey Wansbrough's understanding of the passage in question,
nor his intention in citing it. I would like to express my gratitude to
Gerald Hawting (London), Walid Saleh (Toronto), Eliezer Segal
(Calgary), and Ingrid Holmberg (Victoria) for their help with various
elements of these annotations; I do accept full personal responsibility
for what has been provided here, however.

ANNOTATIONS
Parti

259

Page x

wie es eigentlich
gewesen

Page 2

Q. 2:134

0:7=34

Q, 10:47
Q. 16:63
Q. 23:81
0,8:38

0:6:25

Q. 15:10

0:43:8
Q. 6:6

0,7:4
0:15=4

Q: 6:131

Q. 20:99

how it actually happened

tilka ummatun qad khalat

li-kulli ummatin ajal

li-kulli ummatin rasul

laqad arsalna ila umam min qablika

bal qalu mithla ma qala '1-awwalun

fa-qad madat sunnatu i-awwalin

in hadha ilia asatiru '1-awwalin

wa-laqad arsalna min qablika fi shiya'i '1-

awwalln

wa mada mathalu 1-awwalin
kam ahlakna min qablihim min qarn
wa kam min qaryatin ahlaknaha
wa ma ahlakna min qarya
mublik al-qura

kadhalika naqussu 'alayka min anba' ma qad
sabaqa

Page 3

Q. 12:111
Q. 30:42
Q. 27:34

Ubi sunt qui ante nos
in mundo fuere

laqad kana If qisasihim 'ibrat
fa-nzuru kayfa kana 'aqibat alladhina min qablu
qalat inna '1-muluk idha dakhalu qaryatan
afsaduha wa j'a'alu a'izzat ahliha adhillatan wa
kadhalika yaf aluna, She said, "Kings, when
they enter a city, they disorder it and make the
mighty ones of its inhabitants abased. This is
what they will do."

where are those who lived before us

Page 4

0:13:6
Q. 34:19
Q. 2:49

waqad khalat min qablihim al-mathulat

wa [fa] ja c alnahum ahadlth

wa fi dhalikum bala'min rabbikum

260

ANNOTATIONS

0,89:15

/* >l - _ . 11 _ - 1/ . /-/ 111

la amma 1-msan ldha ma btalanu rabbunu

0; 29:3

1 If — Jll 17 — • J 7*7 '

wa laqad manna Uadhina mm qabhhim

Q 21:35

1 77 /* 11 —}• .31 1 1—1 f *-?/

kullu nars aha lqatu l-mawt wa-nablukum 01 1-

sharr wa'l-khayr fitnatan wa-ilayna turja'iina

0:74:3!

7 77 7-1 I'll 311—1 1 —y

kadhalika yudillu llah man yasha u

Q. 11:116

1 • 1 1'1 — 1— 1 — <

al-qurun mm qablikum ulu baqiyat

0: 30:43

11' y. • 1— J J 1 I

mm qabh an ya tiyayawm la maradd Janu mm

Allah

Q: 32:14

nasi turn hqa yawmikum

0: 34:30

lakum mfad yawm

Q. 69^4

1 — 111 — 7 •

al-ayyam al-khaliya

Q. 10:102

mithla ayyam alladhma khalaw min qablihim

0=45:14

alladhlna [Madhlna] la yarjuna ayyam Allah

Biblical

yom YHWH

Isaiah 9:3

yom Madyan

ludfjes 8:1 3

malhamah

Numbers 21:14

milhamot YHWH

Q. 3:140

. -77 j7 — 1- '1 I ~ 1 i7

wa hlka I-ayyam midawiluha bayna 1-nas

0:14=5

akhrij qawmaka min al-zulumat ilia '1-nur wa

dhakkirhum bi-ayyam Allah

Page 5

Q. 3:190

inna fi khalqi '1-samawat wal-ard wa-'khtilaG '1-

layl wal-nahar la-ayat H ull '1-albab

Nachrichten

apokalyptischer Art communications of an apocalyptic character

Page 6

Q: 38:29

kitab anzalnahu ilayka mubarak li-yaddabbaru

ayatihi

(a 55)

(Surat Rahman): 13

fa-bi'ayyi ala rabbikuma tukadhdhibani

Q, 2:248

inna fi dhalika la-ayatan

Isaiah 37:30

we-zeh lekha ha-ot

Isaiah 7:14

laken yitten adonayhu lakhem ot

0,18:9

am hasibta anna ashab al-kahf wa-'fraqim

kanii min ayatina 'ajab

Heilsankundigung

announcement of salvation

Gerichtsankiindigung

announcement of judgment

ANNOTATIONS

s Page 7

Q. 10:20
0,21:5

O 37:168
Q. 7:203

o 40:78

Q. 19:46, 48

Q. 73:10
Q. 2:218
Q. 4:100

Page 8

0,8:5

0: 59=3

9:38
0: 51:50

O 44:23

O 11.8l
0: 3:137

3:8l

Page 9

Qi 9:75

O 2:27
O 7:169

lawla unzila 'alayhi ay a min rabbihi

bal qalu adghath ahlam bal iftarahu bal huwa

sha'ir falya'tina bi-aya kama ursila '1-awwalun

law anna Indana dhikr min al-awwalin

wa-idha lam ta'tihim bi-aya qalu lawla

jtabaytaha

wa ma kana li-rasul an ya'tiya bi-aya ilia bi-
idhn Allah

qala araghib anta 'an alihatlya Ibrahim lain
lam tantahi la-arjumannaka wa 'hjurni maliy
. . . wa-a'tazilukum wa ma tad'una min duna
'Hah

wa-'sbir 'ala mayaquluna wa-'hjurhum hajr
Jamil

inna 'Uadhlna amanu wa-'lladnina haraju wa
jahadu fi sabil Allah

wa man yuhajir fi sabil Allah yajid fi '1-ard
muraghaman kathiran wa sa 'atan wa man
yakhruj min baytihi muhajiran ila Allah wa
rasulihi

kama akhrajaka rabbuka min baytika bi'l-haqq

wa lawla an kataba 'Hah 'alayhim al-jala' ' la-

'adhdhabahum fi '1-dunya

idha qlla lakum unfirii fi sabil Allah

fa-firru ila 'Hah

la-ash bi-'ibadi laylan

fa-ash bi-ahlika bi-qit' min al-layl

fa-slru fi '1-ard fa-'nzuru kayfa kana 'aqibatu '1-

mukadhdhibm

wa-idh akhadha 'Hah mlthaq al-nablyln

wa minhum man 'ahada 'llah

alladhina yanquduna 'ahda 'llah min ba'd

mlthaqihi

a-lam yu'khadh 'alayhim mlthaq al-kitab

262

ANNOTATIONS

Saadya Psalm 111:5 yazkhar le-ohm berito, "He is ever mindful of
His covenant," rendered in Judaeo- Arabic as
yadhkuruna bihi ila '1-abad 'ahdahu, "they will
remember thereby forever His covenant."

Q. 8:56 alladhma 'ahadta minhum thumma yanquduna

'ahdahum fl kulli marratin wa-hum la
yattaquna

Q. 8:72 wa-in istansarukum fif '1-din fa-'alaykum aJ-nasr

ilia 'ala qawm baynakum wa baynahum mitbaq

Q. 19:78 attala'a '1-ghayb am ittakhadba mda '1-rahman

'ahdan

(^17:34 inna 'l-'abd kana mas'ul

Q. 33:15 wa kana 'ahd Allah mas'ul

Page 10

diatheke

Terminus der
arabischen
Rechtssprache

Umma document

0:23:8

Nehemiah 10:1
Q. 48:10

Q. 16:91

covenant

terminology of Arabic jurisprudence
wa-inna dhimmat Allah wahida
wa-'lladhlna hum li-amanatihim wa-'ahdihim
ra'una
amanah

inna lladhma yubayi'unaka innama yubayi'una
'llah

wa-awfu bi-'ahd Allah idha 'ahadtum wa la
tanqudu 1-ayman bada tawkldiha

Page 11

0; 33^5
Q: 337

Q. 20:115

Jeremiah 31 :3i
Luke 22:20

wa laqad kanu 'ahadu 'llah min qablu la
yuwalluna '1-adbar wa kana 'ahd Allah mas'ul
wa-idh akhadhna min al-nablyin mithaqahum
wa minka wa min Nuh wa-Ibrahlm wa-Musa
wa-'Isa ibn Maryam wa-akhadhna minhum
mithaq ghallz

wa-laqad 'ahidna ila Adam win qablu fa-nasiya

wa lam najid lahu azm

bent hadashah, "new covenant"

kaine diatheke, "new covenant"

ANNOTATIONS

263

ij Q. 9:1 bara'a min Allah wa rasulihi ila 'lladhina

'ahadtum min al-mushrikln
Q. 54:43 a-kuffarukum khayr min ula'ikum am lakum

bara'a fi 1-zubur

Page 12

Zamakhshari

Q. 19:17
Q. 19:19
Q. 19:21

Page 13

(0=19:21)
Hebrew

Q. 10:33

Q. 27:34

Q. 11:102

Q. 18:19
Q. 2:286

0:4=47

Q. 2:187
0=6:55

Q: 37:80

prophetisches Wort
Botenwort

Concerning the decrees and orders of the
kings, the following (expressions) are used: the
king has directed {taqaddama) someone,
instructed (aw'aza) him, commanded ( azama)
him, and covenanted ( 'ahida) with him.
fa-arsalna ilayha ruhana fa-tamaththala laha
bashar sawly

qala innama ana rasul rabbika li-ahaba laki
ghulam zakiy

qala kadhaliki (the final vowel reads so in the
Qur'an, although the standard would be
kadhalika) qala rabbuki huwa 'alayya hayyin
wa li-naj'alahu ay a lil-nas wa rahma minna wa
kana amr maqdiy

kadhalika qala rabbuki
koh amar YHWH

kadhalika haqqat kalimat Allah [rabbika]
wa kadhalika yafaluna

wa kadhalika akhdh rabbika idha akhadha 1-
qura

wa kadhalika ba'athnahum

kama hamaltahu 'ala 'lladhina min qablina

kama la'anna ashab al-sabt

kadhalika yubayyinu llah ayatahu

wa-kadhalika nufassilu 1-ayat

inna kadhalika najzi l-muhsinin

prophetic speech as the speech of a messenger

Page 15

Ezekiel 3:1 1 Speak to them and say to them, "Thus says the

Lord God."

264

Ezekiel 1 1:5
(Proverbs 5:7)
(Proverbs 4:10)
(Proverbs 4:1)

Page 16

Q, 31:12-13

Ezekiel 8:6
Ezekiel 8:17
Kitab al-asnam

Page-17
Muslim talbiya
Deuteronomy 33:29

Page 18

0; 3735

Q. 29:45

Q. 112 (Surat Ikhlas)

Q. 6:121

Hebrew

ANNOTATIONS

He said to me, Say, "Thus says the Lord.'
we- c atah banim (Listen to me!)
shema ' beni
shim'u banim

wa-laqad atayna Luqman al-hikma . . . wa-idh

qala Luqman li-'bnihi wa huwa ya'izuhu ya

bunayya la tushrik bi-'llah

Son of man {ben adani): do you see?

Have you seen, son of man (ben adani) 7 .

with their mixing into that which was not of it

labbayka

Happy are you! (ashrayka)

la ilah ilia llah

[wa-dhkuru] ism Allah

wa-ladhikr Allah akbaru

qui huwa llah ahad

minima lam yudhkar ism Allah 'aJayhi

we-'atah

Page 19

Q. 2:248
Q. 20:9

0:970

Q: 7:l85
Q.34:19

a 18:13

Q. 12:3
Q. 11:49

0:4:44
Q. 13:41
0.30:9
Q. 19:16

inna fi dhalika la-aya

hal ataka hadlth [Q. 38:21 reads hal ataka

naba'u\

a-lam [ya'tihimj naba'lQ. !4 : 9 — not 14:19-

reads adam [ya'tikumj naba]

fa bi-ayy hadlth ba'dahu yu'minuna

[fa] ja'alnahum ahadith

nahnu naqussu 'alayka naba'ahum

nahnu naqussu alayka ahsan al-qasas

tilka/dhalika min anba'

a-lam tara ila/an

aw-lam yaraw ila/an

aw-lamyasiru fi '1-ard

wa-'dhkur ft '1-kitab

ANNOTATIONS

265

S 0:38:17
0:14:5

Page 20

Q: 18:32

0:5:27
Q. 2:252

Page 21

Moubarac

Page 24

Q. 29:36-37

diatribe (Q. 11:84)
Page 25

(0: 11:85)
(Q: H:85)

Q. 2:60

0: 55:46

Q, 55:46-61

Page 26

Psalm 136

Page 27

Q. 2:25

Waraqa b. Nawfal
'Abld b. Abras

wa-'dhkur 'abdana

wa dhakkirhum bi-ayyam Allah

wa-'drib lahum mathal
wa-'tlu 'alayhim naba'
tilka ayat Allah natluha

Moreover, there is a history of the religion in
the Qur'an more than a religion of history.

And to . . . their brother ... he said, "O

people, serve God."

Give full measure and weight

Do not cheat the people of their property

Do not work evil in the land

Do not act corruptly in the land

(janna is singular in Q. 79:41) For he who

fears standing before his Lord, two gardens

(jannatan) — al-Farra° said that this means one

garden (Janna); it is as if He had said, "Indeed,

the garden is the dwelling place," but He made

it dual on account of the rhyme.

And which of the favors of your Lord will the

two of you deny?

For His steadfast love endures forever

Gardens (jannat) under which flow rivers
Among the people he was a tyrant; to the Fire,
he was Hawiya.

I am informing you that I entered the hawiya.

266

Page 28

a 16:51

Page 30

Zamakhshari

Page 31

Q. 18:48

Q. 20:55

Q. 36:79
Q. 10:34

Deuteronomy 32:39
Q. 2:259
liturgical Hebrew

Page 32

Ezekiel 37:1—14

[1. 2, 3]
0.23:14

Q: 3678

Q. 22:5
Q: 56[ 5 8]:6
0,7:14

a 19:85

Q. 10:28
Q. 24:64

0; 19:15/33

Q. 22:7
Q. 30:56
Q. 50:44

r / 1

ANNOTATIONS '

translated in the text

These are the angelic archangels/ cherubim who
surround the throne, like Gabriel, Michael,
Israfll and others of their rank.

You have come to us as We created you the
first time

We created you from it and into it We shall
return you and from it We shall bring you out
again

Say, "He who gave life to them the first time
will give life to them."

Say, "God originates creation and then repeats
it."

I give death and life
I give life and death
Giver of life to the dead

'atsamot yebeshot, "dry bones"
We made bones out of the tissue and clothed
the bones with flesh

Who will give life to the bones when they have
decayed?

Indeed, We created you of dust
yawm yab'athuhum Allah
ila ya wm yub athuna
yawm nahshuru '1-muttaqin
wa-yawm nahshuruhum jam?
wa-yawm yurja'una ilayhi
wa-yawm yub'athu/ub'athu hayy
wa-anna 'llah yab'athu man fi 1-qubur
fa-hadha yawm al-ba'th
yawm tashaqqaqu '1-ard anhum sira' dhalika
hashr 'alayna yaslr

\

S Q 86:8-9

ANNOTATIONS 267
innahu 'ala raj'ihi la-qadir yawm tubla '1-sara'ir

Page 33

Q: 39:68

Q;8 3 :6

anastasis

Hebrew

Greek

fa-idha hum qiyam yanzuruna

yawm yaqumu '1-nas li-rabbi 'l-'alamm

resurrection

Giver of life to the dead

anastasis

Page 34

0.42:51

Q. 42:52

wa ma kana li-bashar / an yukallimahu '11 ah ilia
/ (A) wahy / (B) aw min wara'hi/ab/ (C) aw
yursila rasul / fa-yiihiya bi-idhnihi / ma yasha'u
/ innahu 'ally hakim

Therefore We revealed (awhayna) to you a
spirit from Our command.

Page 35

Biblical masveh/

halymma
Rabbinic Hebrew

[Gen. Rabba 52, 7] me-ahore ha-wilon, "from behind the veil'

Hebrew and Greek respectively: veil

Q. 4:164

Page 36

Q. 25:32

Kalbi on Q; 17:2
Q. 17:106

Q 56:75

Page 37

Q; 20:114
Q2:l85

wa kallama 'Hah Musa takhman

wa qala 'Uadhina kafaru law-la nuzzila 'alayhi 1-
qur'an jumlatan wahidatan kadhalika li-
nuthabbita bihi fu'adaka wa-rattalnahu tartllan
wa-atayna Musa '1-kitab — We bestowed on
Moses the Torah all of one piece (jumla
wahida)

wa qur'anan [faraqnahu] '1-taqra'ahu 'ala 1-nas
'ala mukth wa nazzalnahu tanzllan
bi-mawaqi'i '1-nujum

wa la ta'jal bi'l-qur'an min qabli an tuqda ilayka
wahyuhu

The month of Ramadan in which the Qur'an
was sent down

268

0,97=1
SuyutT

Ibn Ishaq

Saadya

Page 38
SuyutI

Page 39
1.

Q. 11:29

Q: 3:154

Q. 62:2

2.

Page 40

(2)

(2 cont'd)

3-
4-
5-

ANNOTATIONS

Indeed, We sent it down on the night of power
(laylat al-qadr).

It was sent down to the nearest heaven on the
night of power all of one piece then it was sent
down after that serially over the period of
twenty years.

It is within that which has reached me about
what Jesus, son of Mary, had written in what
came to him from God in the Gospel (which
he gave) to the people of the Gospel.
The Torah was given (to Israel) scroll by scroll.

Concerning (this topic) also is the equality
(taswiya) between our prophet and Moses in
the revelation of his book as a whole and the
elevation ( tafdll) of Muhammad in the
revelation to him being serial (munajjam) so
that he could memorize it.

We were a people in ignorance {jahiliyya)
I see you are a people that is ignorant
( tajhaluna)

Until God sent to us a messenger from among
us

It is He who sent the common people a
messenger from among them
Worshiping idols

who called us to God, to associate no one with
him, to worship Him, to rid ourselves of the
stones and images which we and our fathers
had been worshiping

He commanded us to worship God alone and

to associate nothing with Him

and eating carrion

frequenting prostitutes

and severing the ties of kinship

ANNOTATIONS 269

S 6. p'war, security

7. the powerful among us oppressing the weak
Q. 4:10 consuming the property of orphans

8. speak honestly

Q. 19:50 a tongue of truthfulness

Q. 22:30 false speech

(8.) slander of honorable women

Q. 24: 4 those who accuse honorable women

9. prayer and almsgiving; fasting

Page 41

Ja'far first part of Kaf-Ha'-Ya-'Ayn-Sad {another

name of surat Maryam, chapter 19)

Najashl translated in the text

Ja'far translated in the text

Page 42

Hebrew translated in the text

Page 43

Noldeke-Schwally The emergence of the Muslim canon is

completely exceptional; one could even say that
it was formed in opposition to the norm. It is
not the work of several writers, but only of one
man who accomplished it in the short span of a
generation.

Page 46

Arabic translated in the text

Page 49

Kitab al-Mabanl The messenger of God said, "I was given the

seven long (chapters) in place of the Torah and
the hundred (chapters) in the place of the
Gospel and the mathani in the place of the
Psalms. And I was honored by their
arrangement."

270

Page 51

Tabari

Juwayni

ANNOTATIONS

The Qur'an was in many books, and you
abandoned all of them but one.
I say that the Qur'an is one part and the other
part is the sunna, just as it is said that Gabriel
has revealed the sunna just as he revealed the
Qur'an. From this comes the transmission of
the sunna according to the meaning because
Gabriel brought it in its meaning (only). But
the Qur'an did not come in its meaning (only)
because Gabriel gave it in its wording.

Part II

Page 53

Q. 14:4

Q 10:47
Qi3:8

Q: 35^4

Q 28:75
Q. 16:113

Numbers Rabba
2o[, 1]

We do not send a messenger unless (he comes)

in the language of his people.

Every community has a messenger.

Every people has a guide.

There has not existed a nation without a

warner having passed away in it.

We shall draw out from every nation a witness.

There came to them a messenger from among

them but they cried lies against him.

In the same way as He raised up kings, sages,
and prophets for Israel, so He raised them up
for the ummot ha-'olam.

Page 54
Q7:i57-58

am ha-ares
0:3:67

Q. 16:120
Genesis 12:2

the messenger, the unlettered prophet (or: the
messenger, the prophet of the common folk)
common folk

Abraham was not a Jew nor a Christian but he
was a man of pure faith (hanlf), a Muslim.
Abraham was a nation obedient to God, a man
of pure faith (hanlf).
And I will make of you a great nation

ANNOTATIONS

) Page 55
Malachi 3:1

O 2:285
Q.4l:43
0,46:9

Q: 17:55
4:125
Isaiah 41:8

Page 56

ausserer

Geschichtsablauf

innere Einstellung

Page 57

Suyuti

imitatio magistri
al-asl al-iqtida'
magister dixit
Hebrew
symbolic acts
sayings

explication de texte
Page 58

Grundschicht, erste
Schicht, zweite
Schicht

Behold, I send my messenger (malaki)

We make no division between any one of His

messengers.

Nothing has been said to you without it having
been said to the messengers before you.
Say, "I am not an innovation among the
messengers."

We have preferred some prophets over others.
God took Abraham as a friend
Abraham, my friend

historical processes taking place around
someone

inner, personal attitudes

He said, may the prayers and peace of God be
upon him, "Was I not given the Qur'an and
the like of it with it?" that is, the sunna.
imitation of the master
principle of emulation
"the master said it"
The statutes given to Moses on Sinai.
Arabic: sunna fa'liyya; Hebrew: ma'asim
Arabic: sunna qawliyya; Hebrew: debarim
close reading of a text

basic (foundational) stratum, first stratum,
second stratum

Page 59

0,91:8

and inspired {alhama) it to lewdness and
godfearing

272 ANNOTATIONS

Kalbl ac/Q. 12:15 We inspired (awhayna ila) him — (that is,)

Joseph; He sent (arsala) Gabriel to him. It is
also said, He inspired him {alhamahu).
Suyuti God inspired (aJhama) the righteous caliphs

Ibn Qudama Then it is necessary for you to say that poetry

is a qur'an because God inspires (alhama) the
speech of the poets and empowers them, just as
He inspires (alhama) Gabriel according to your
statement.

Suyuti The prophet was miserable and did not arise

for a night or two. His wife came to him and
said, "O Muhammad, I can only think that
your Satan has left you."

Page 60

Qj 69:41 It is not the speech of a poet.

Q. 69:42 Nor the speech of a soothsayer

Q: 37-36 What, shall we forsake our gods for a poet

possessed?

Q. 51 :52 Even so, there was not a messenger who came

to those before them without them saying, "A
sorcerer or a man possessed."

Q. 40:15 He casts (yulqt) the spirit of His bidding.

Q. 20:39 And I cast (alqaytu) love on you from Me.

Q. 8:12 I will cast terror (uJqf) into the hearts of those

who disbelieve.

Suyuti or it was preserved in the heavenly tablet, so

(Gabriel) came down to the messenger and cast
it (yulqlhi) into him.

Q. 22:52 Satan casts (alqa) into his desire ... so God

abrogates what Satan casts {yulqt).

Page 61

Q. 6:1 1 2 So We have appointed to every prophet (nabi)

an enemy, satans of men and jinn revealing
adorned speech to one another.

Q. 72: 27—28 Only of those messengers with whom He is

well pleased; then He dispatches guarding
angels (rasad) before him and behind him so

ANNOTATIONS

273

Zamakhshari

Page 62

Kalbl on Q. 12:2

(a 2:97)

Page 63

(gloss of umml in

0:7:157-58)
'am ha-ares
anthropoid

agrammatoi
(ac/Q, 7:157)

Page 64

Hebrew

Q. 61:6
0; 33:40

seal of the prophets
(place a seal upon)
Haggai 2:23

that He may know they have delivered the
messages of their Lord.

Protecting guardians (rasad) among the angels
protect him from the satans.

We revealed it as an Arabic Qur'an — He is
saying, We sent Gabriel down to Muhammad
with the Qur'an following the rules of the
Arabic language.

He has brought it {hu, accusative pronoun)
down on your heart by the permission of God.

He who could not write or read
common people

uneducated people

The illiterate messenger prophet about whom
they find written reference in their Torah and
Gospel

Notriqon (deriving meaning by breaking up a
word into two or more words or into individual
letters and having each letter stand for another
word starting with that letter) and Gematria
(computation of the numeric value of letters; or
substitution of letters for other letters).
Good tidings of a messenger who shall come
after me whose name shall be Ahmad.
Muhammad is not the father of any one of
your men but the messenger of God and the
seal of the prophets,
last of the prophets
khatam : taba'
hotem

274

ANNOTATIONS

Page 65

I Corinthians 9:2
Romans 10:4
Greek

Page 67

Q, 94:i-3

spiritual vision

0: 17:i

Page 68

(0, 17:1)

(a 17:1)

(0: 17:1)

Q. 17:60
Q. 17:93
Page 69

Muqatil adQ. 17:1

Page 71
Ibn Hisham

sphragis, "seal"

telos nomou, "goal of the law"
theios anthropos, "godlike man"

Did We not open your breast for you and lift
your burden from you that weighed down your
back?

Arabic: a vision in the heart; Hebrew: vision of
the heart

Glory be to Him who carried His servant by
night from the holy mosque to the further
mosque.

carried His servant ( abet) by night {laylan)
Glory be to He who

from the sacred mosque to the distant mosque
We made the vision that We showed you to be
only a trial for men.
or you ascend up into the heaven

Only journey to the three mosques.

I am first to revive the order of God and His
book and to practice it.

Page 72

Q. 22:30 And permitted to you are the flocks except that

which is recited to you. Avoid the abomination
of idols.

Q. 17:45 When you recite the Qur'an We place between

you and those who do not believe in the world
to come an obscuring curtain.

Q; 5:13 We made their hearts hard.

Q; 2:88 And they say, "Our hearts are uncircumcised."

ANNOTATIONS

275

\ Leviticus 26:41 uncircumcised heart (often translated as

"stubborn spirit ")

Page 73

Exodus 4:21 I will harden his heart (often translated as

"make him obstinate").
Exodus 34:33—35 masveh
2 Corinthians 3:12— 18 kalumma, "veil"
Tabari ad Q. 17:45 translated in the text
Psalm 88:15

[= 88:14 KJV etc.] hiding Your face from me
Saadya, Psalm 88:1 5 Do not veil your mercy.

Page 74

Q. 7:85 A sign has come to you from your Lord.

Q. 4:153 The people of the book will ask you to bring

down a book from heaven.

Page 75

Q. 29:27 prophethood and the book (kitab)

Q. 3:79 the book (kitab), wisdom, and prophethood

Q. 35:40 Or have We given them a book so that they

have a sign from it?
Tabari I found it predicted in the book of God (may

be He praised and elevated), the Torah.

Page 76

Q. 2:285 Each one believes in God and His angels and

His books (kutubihi) and His messengers.

Q. 4:46 Some of the Jews change (yuharriftina) the

words (kalim) from their meanings.

Q. 50:29 The word (qawl) is not changed with Me.

Q. 6:1 15 None can change His words (kalimat).

Q. 48:15 They desire to change God's speech (kalam).

Page 77

Hebrew
Psalm 33:9

debar YHWH, "word of God"

For he spoke, and it was; he commanded, and

it stood firm.

2j6

Q. 2:1 17

ANNOTATIONS

And when He decrees a thing, He says to it,
"Be," and it is.

Page 79

Q. 17:88

Q. 28:49

Page 83

0,9=6

Q. 85:21—22

Page 84

awir sheni

Say, "If men and jinn joined together to
produce the like of this Qur'an, they would not
produce its like."

Say, "Bring a book from God that gives better
guidance than these two."

Grant him protection until he hears the speech
of God

Rather it is a glorious Qur'an in a guarded
(mah&z) tablet

"second air" (that being the spirit of God from
which all things emanate, interpreted as a
subde, rarefied air in Saadya's Commentary on
the Sefer Yesirah.)

Part III

Page 88

le fait coranique

Page 93

hadith

Page 94

Basran grammarians

the Qur'anic accomplishment

I am the most eloquent of the Arabs
although/and I am from Quraysh.

We took the dialect (information) from the
hunters of lizards.

Page 96

Greek

arete, "excellence"

ANNOTATIONS

277

Page 98

Q. 14:4 We do not send a messenger unless (he comes)

in the language of his people.
Q. 4 1 :44 And if We had made this a foreign Qur'an they

would have said, "Why are its verses not

distinguished? Is this a foreign (book) (with)

an Arab?"

Q. 16:103 Indeed, We know that they say, "Only a mortal

is teaching him." The speech of him at whom
they hint is foreign and this is clear Arabic
speech.

Page 99

Q. 26:194—95 Upon your heart so that you may be one of the

warners in a clear Arabic speech.

Zamakhshari

on Q. 13:37 an Ar a bic judgment — translated into the

language of the Arabs
Q. 20:27 Unloose the knot upon my tongue.

Q. 28:34 ( ne is) more eloquent than me in speech

Q. 19:97 We have made it easy by your speech.

Q. 55:17 We have made the Qur'an easy.

Exodus 4:15 I will be in your mouth.

Page 100

Aramaic
Abgerissenheit
Stilmischung
Vielsch ichtigkeit
Hintergriindigkeit

bar anash
disjointedness
mixing of styles
multilayeredness
enigmaticity

Page 101

Basran formula This is a mere claim for which there is no

evidence except from wahy and tanzil.

Page 102

gehobene Sprache/

Verheissungsstil elevated language/oracular style
tabellarische Ubersicht tabulated synopsis

278 ANNOTATIONS
Page 103

Q. 2:31 He taught Adam all of the names.

Suyuti He enabled Adam to bestow (the names) on

them.

Suyuti We do not know of a language which has

emerged after him (Muhammad).

Page 104

Ezekiel 3:5 Not to a people whose speech is thick and

difficult have you been sent but to the house of
Israel.

Q. 14:4 We do not send a messenger unless (he comes)

in the language of his people.
Suyuti, hadith translated in the text

Tustarl Recite the Qur'an with the rhythmic

embellishments of the Arabs without
burdening it with other things. Do not read it
with the embellishments of the people of the
churches and the synagogues, nor the people of
heresy and innovation. I and my god-fearing
community are relieved of (all) burdens. People
after me will begin to return their voices to
those of the singers with melodies, captivating
their hearts by captivating the listening heart.
Those are the heedless.

Page 105

Jahill poets Until Iblis ruled their hearts just as he ruled

the hearts of the poets of the jahiliyya
Abdallah translated in the text

Page 107

Ibn Qutayba This is the killer of my brother / This is a

killer, my brother.
Ibn Qutayba A Qurayshi will not be killed / Let him not kill

a Qurayshi bound head and foot after today.
Q. 36:76 Let not their speech grieve you. Indeed We /

such that We know what they conceal and

what they display.

ANNOTATIONS

279

S Page 108

Zamakhshari li-anna
Q. 2:124

0:9=3

a 35:28

Suyuti

Q. 20:63
0,1:4

Q. 41:17

0: 15:54

Q, 39:64

Page 109

piece justificative
Suyuti

Vollers
Wetzstein

with the omission of the lam of explanation
And remember when his Lord tried Abraham
(compare: "when Abraham tried his Lord," the
more natural reading).

that God and His messenger are free from the
polytheists (compare: "that God is free from
the polytheists and His messenger," the more
natural reading).

Rather the learned ones among his servants
who fear God (compare: "rather God fears the
learned ones among his servants," the more
natural reading).

What can be read three ways in the wording of

the Qur'an

in hathani la-sahirani

iyyaka tu'badu (You alone are worshiped);

iyyaka na'budu (You alone we worship)

amma Thamudu/Thamuda fa-hadaynahum

fa-bima tubashshiruna/tubashshiruni

qui afa-ghayr Allah ta'murunnl/ta'murunanl

a'budu ayyaha al-jahiluna

supporting ("justifying") document

The intention of (the recitation of the Qur'an's)

i'rab is the knowledge of the meaning of its

words. The intention of it is not the i'rab as it

is generally accepted among the grammarians

which approaches solecism (lahn) because

reading in the absence (of the full i'rab) is not a

reading at all, and it has no merit.

the stylistic and rhetorical refinements (that

follow from the grammatical i'rab)

All of these points concerning the omission or

curtailing of the short vowels hold for

colloquial language but not for poetry. The

poetry of the Bedouin is certainly to be sung,

and, while the song does not thrive on the

28o

ANNOTATIONS

contrast between consonants, it does need
vowels.

Page no

Derenbourg

Page in

Suyuti

Moreover, all Arab poetry, which had been
transmitted for a long time only by oral
tradition, would lose its rhythm, an essential
attribute of it, if one does not understand the
authenticity of declension in Arabic.

Those who read the Quran who do so with
i'rab will receive twenty blessings for each letter
whereas those who recite it without i'rab will
receive ten for each.

Page 112

Noldeke

prasentisch resultativ
Page 113

disjunctive formulae
Q. 3:117

"uncertainty," "clumsiness," "embarrassing
difficulties," "repetitions for emphasis," "to the
Arabs, nearly everything was new," "every
beginning is difficult"
present resultative (as a verb tense)

God is all-Mighty, all- Wise.

God wronged them not, but they wronged

themselves.

Page 114

Zamakhshari
on Q. 31:28

Q, 27:1
Q:4:i36
Q. 2:118
Q. 6:161

(It is) as if one thing distracted Him from

another, or one action (distracted Him) from

doing something else.

verses of the Qur'an and a clear book

O you who believe

from before them

As for me, my Lord has guided me to a straight
path, a right religion, the creed of Abraham, a
man of pure faith; he was not from among the
polytheists.

ANNOTATIONS

Q. 12:47
Q. 12:74-75
Q. 12:17

except a little of which you may eat
It is its recompense.

You would not believe us though we speak
truthfully.

Page 115

stilistische

Eigentumlichkeiten stylistic peculiarities

Q. 6:109
0,7:2

What will make you perceive that, when it

comes, they will not believe?

A book sent down to you. So let there be no

impediment in your breast because of it, to

warn thereby, and as a reminder to the

believers.

Page 116

Noldeke

The Qur'an created a (category of) literature
for itself; it had no real predecessor and could
have no successor.

Page 117

Genesis 17:20

I have blessed him, I have made him fruitful,
and I have increased him greatly. (This is
translated in the RSV in the future tense, "I
will bless him and make him fruitful and
multiply him exceedingly.")

Part IV

Page 123

Muqatil adQ. 18:4

Page 124

Muqatil

adQ, 18:108-10

Those who say, "God has taken a son" — that
is, Quraysh when they said, "We worship the
angels who are the daughters of God."

Dwelling therein forever (khalidln fiha) — th ey
will not die (la yamutun); they will not desire
any change to them — that is (ya'ni), changing
to something else. That is that ( wa-dhalika

282

ANNOTATIONS

anna) the Jews said to the prophet, "You claim
that you have been given wisdom and that
wisdom is all of knowledge; yet you claim that
you have no knowledge of the spirit and you
claim the spirit is from the command of my
Lord (Q. 17:85). So how can this be?" God,
may His mention be exalted, said to His
prophet, "You have been given knowledge and
your knowledge is only a small part of the
knowledge of God." So (God), may He be
praised {& qala subhanahu), said to the Jews,
Say, "If the oceans were ink for the words of
my Lord the sea would be spent before the
words of my Lord are spent."

Page 125

Ibn Ishaq ad Q. 17:85 Ibn Ishaq said, "I was informed on the

authority of Ibn 'Abbas that he said that when
the messenger of God came to Medina, the
Jewish rabbis {ahbaryahud) said, 'When you
said, And you have only a little knowledge
about it, did you mean us or your own people?'
He said, 'Both of you.' They said, Yet you will
read in what you brought that we were given
the Torah in which there is an explanation of
everything.' The messenger of God said, 'Of
God's knowledge there was little in it, but
among you there is sufficient for you
concerning that, if you carried it out.'"

Suyuft ad Q. 2:97 He said that 'Abd Allah ibn Salam heard of the

coming of the messenger of God, so he went to
him. He said, "I will ask you about three things
of which only a prophet has knowledge. What
are the signs of the day of judgment? What is
the food of the people of paradise? What will
the child take over from his father or his
mother?" (Muhammad) said, "Gabriel will
inform me of these things." Gabriel said, "Yes."
He said, "That is the enemy of the Jews from
the angels." So he recited this verse (Q. 2:97).

ANNOTATIONS

Page 126

Noldeke

Noldeke
Noldeke

283

The entire matter is so fantastic that we have

very little to go on.

either from Mecca or from Medina

We have one significant advantage over them:

impartiality rather than religious prejudice. As

well, we have been raised in the school of

scientific criticism.

Page 129

Muqatil ad Q. 18:45 (God), may He be praised, is saying that the
likeness of this world is the likeness of the
vegetation when it becomes green which then
dries out and dies. Similarly, the world will be
destroyed when the end of time comes.

Muqatil at the end of

surat al-Kahf (18) God is saying, "1 am the best of partners.

Whoever partners with Me in deeds, I will
create all (of those) deeds for my partner. I
only turn towards what is Mine purely."

Page 131

Kalbi on Q. 12:56—57 So We establi s hed Joseph — Thus We made

Joseph ruler; in the land — the land of Egypt; to
make his dwelling there — settling in it;
wherever he would — he desires. We visit with
Our mercy — We make Our mercy (of)
prophethood and Islam appear; whomsoever
We will — to whoever is qualified for that; and
We leave not to waste — We do not corrupt; the
wage of t he good-doers — the reward of the
righteous believers in word or deed; and the
reward of the world to come — recompense
[read thawab] of the world to come; is better —
than the recompense of the world; for t hose
who believe — in God and all the books and
messengers; and are god-fearing — (of) disbelief,
association and corruption.

284 ANNOTATIONS
Page 132

Kalbi adQ. 12:6 And thus — therefore; Your Lord will choose

you — Your Lord will select you for
prophethood; and teach you the interpretation
of dreams — explanation of visions; and will
perfect His blessing on you — with prophethood
and Islam; that is (ay), He will cause you to die
in that state; and on the family of Jacob — with
you, that is (ay) He also perfects His blessing
on the children of Jacob along with you; just as
He perfected it — His blessing of prophethood
and Islam; on your fathers before - — before you
[reading qablaka as in printed text]; Abraham
and Isaac; Indeed God is all-Knowing — of His
blessing; all- Wise — of its perfection

Kalbi ad Q. 12:63 So send our brother with us — Benjamin; We

will obtain the measure (naktal) — he will
obtain a load for himself, and it is said, we will
obtain a load for him, if (the text) is read with
a nun.

Page 133

Farra" ad Q. 12:63 The followers of 'Abd Allah read yaktal while

all the others read naktal. Both are correct.
Those who say naktal include him with them
in the measure whereas those who say yaktal
attribute a load to him by himself. So the
(load) is attributed to him singly because they
increase the (overall) measure by a camel load
with his (load).

Page 134

Kalbi adQ. 12:43 translated in the text
Genesis 41:2 they came up from the river

Genesis 37:3 long-sleeved robe

Speyer The mention of the "largest of them," here

kabir, surely means "the eldest," thus Rubin.

\

S Page 136

Kalbl adQ. I2:z4

Genesis Rabba 87,
9 [?]

Kalbl adQ. 12:67
Genesis Rabba 91,2

Page 138

Kalbl adQ. 12:31

Sufyan adQ. 12:31

Page 143

Kalbl adQ. 12:87

Muqatil adQ 18:22

ANNOTATIONS 285

had he not seen the proof of his Lord — a
compelling warning from his Lord to him; and
it is said he saw the image of his father. It is
also said, had he not seen the proof of his
Lord — due to the significance of it.

He saw his father's face, at which his blood
cooled.

Jacob feared for them because of the (evil) eye.
Do not all enter through one gate. Do not go
out with bread in your hands so as not to
arouse ill-feeling, and do not all enter through
one gate for fear of the (evil) eye.

She prepared for them muttaka'an — pillows on
which they reclined, if one reads this with a
shadda; if one reads it without shadda (thus
reading mutkari), He is saying, citrus.
Sufyan told us on the authority of Mansur
from Mujahid that he said that he who recites
it as muttaka'an with a nun says it is a meal,
while those who read it without the nun say it
is a citrus fruit.

Do not despair of the spirit of God — of the
mercy of God.

It begins with a waw because that marks the
end of (a section of) speech. Abu '1-' Abbas
Tha'lab said that al-Farra' said that this waw is
the waw of the hal (i.e., indicating "while" or
"when") and the meaning is that this is their
condition at the time of the mention of the
dog.

Page 145

Muqatil adQ. 18:69

He said, "You will find me, if God wills,

patient" — Muqatil said that Moses was not

2 86

ANNOTATIONS

patient and was not influenced by his
statement.

Kalbl adQ 12:99 and he said , " Enter Egypt, if God wills, in

safety^ — God had willed (them) safety from
(their enemies) and evil.

Page 147

predication coranique Qur'anic preaching
Page 148

Arabic from Goldziher a weak hadith but one feels content with it

Page 149

Q. 3:7 It is He who sent down upon you the book

wherein are clear verses {ayat muhkamai) that
are the essence of the book ( umm al-kitab) and
others ambiguous {mutashabihat). As for those
in whose hearts is swerving, they follow the
ambiguous part, desiring dissension and
desiring its interpretation. And none know its
interpretation except God. And those firmly
rooted in knowledge say, "We believe in it; all
is from our Lord." But none remember other
than those possessed of minds.

Four initial sigla alif-lam-mim wa alif-lam-mlm-sad wa alif-lam-

mlm-ra' wa alif-lam-ra'

Page 150

Abu 'Ubayd on Q. 3:7 The muhkamat are (the Qur'an's) abrogating

verses and the statements of the permitted and
the forbidden and the religious duties that are
believed in and acted upon. The mutashabihat
are abrogated verses, verses displaying
inversion, parables, oaths, and verses that are
believed in but not acted upon.

Page 151

Maturldi on Q. 3:7 It is said that it is possible that one may

understand the mutashabihat on the basis of
knowledge of the muhkamat.

ANNOTATIONS

■ s Jassas on Q. 3:7

Ibn Qutayba

287

It is established thereby that the intention
behind mentioning the mutashabih in this
verse is (the idea) that the sense is connected to
the meaning which is the opposite of the
muhkam and that fixes its meaning.
If all of the Qur'an was obvious and open such
that both the intelligent and the ignorant were
equally able to understand it, this would mean
the loss of the differentiation between people,
the ceasing of the trials (of life), and the death
of (significant) thought.

Page 152

Suyuti

Bacher
Bacher

such that some of the mutashabih pertain
specifically-txrGod, most High
The Torah speaks in the language of humans.
A text may not be distorted from its obvious
meaning {peshut).

Page 153

Q: 13:39
Q: 43:4

Rabbinic precepts

Q: 75:i9

Allah . . . wa-'indahu umm al-kitab
wa-innahu fi umm al-kitab ladayna
yesh em lammiqra' / lamasoret — literally:
"there is a mother to the reading/ tradition,"
thus meaning, "the (vocalized) reading /
written (consonantal) text is primary" (this is a
debate within Rabbinic texts over the textual
basis on which exegesis should be conducted)
thumma inna 'alayna bayanahu

Page 154

0; 25:33

Matundi

They do not come to you (ya'tunaka) with a
similitude unless We bring you the truth and
better in exposition ( tafslr)
translated in the text

Page 155

Maturidi

tafsirh&s only one aspect, ta'ml has many
aspects

288

SuyutT

Page 157
'Abdallah

Page 159

0:7:43

Ibn Muqaffa'
Ibn Muqaffa'

Page 161

Hasan Basri

Hasan Basri

ANNOTATIONS

The Qur'an elucidates many different aspects
( wuj'uh) I the Qur'an conveys many different
aspects ( wujuh).

We have killed you with its
interpretation/outcome {ta'wlt) I just as we
have killed you with its revelation.

We would not have been guided if God had
not guided us.

according to the agreement of the messenger of
God and the rightly guided leaders (successors)
after him

So he will pay attention in it to the most
truthful of the two parties and to the most
equitable comparison of the two cases.

Is it on the authority of a transmission of one
of the companions of the messenger of God or
is it according to an opinion of his or is it an
argument the truthfulness of which is
confirmed in the Qur'an?
We have learned (this), O commander of the
faithful (may God make you thrive!), from the
pious ancestors who followed the command of
God and considered His wisdom and followed
the sunna of the messenger of God. They did
not reject the truth, nor did they accept
falsehood as truth; they did not ascribe to the
Lord more than that which He allowed
Himself, nor did they use any arguments other
than those used by God in addressing His
creation in His book.

Page 162

Hasan Basri

He knows of its soundness according to the
book of God and the sunna of the messenger

ANNOTATIONS

289

Hasan Basri
Hasan Basri
Hasan Basri

Hasan Basri
on Q. 19:59

of God . . . , so know that there is, among
those who transmit on the authority of the
ancient pious ancestors from the companions
of the messenger of God, no one who is more
knowledgeable about God most high and more
learned in religion and better versed in the
book of God than al-Hasan.
translated in the text
translated in the text

They interpret {yata'awwiluna) that according
to their opinion / They interpret {yufassiruna)
that according to their opinion.

That is, a great punishment; the Arabs say,
"So-and-so encountered an offense today,"
meaning that the commander struck him with
a severe blow and punished him with a great
punishment.

Page 163

Hasan Basri
on Q. 3:178

Hasan Basri

The Qur'an, O commander of the faithful, is in
Arabic. God revealed it to the community of
Arabs, addressing them in their speech, the
meaning of which they know.
The people opposed the book of God, O
commander of the faithful, and they altered it
and distorted some it into what it was not.
Rather, it was just as the best of narratives
describes a book in which some parts of it are
analogous {mutashabih) to other parts of it and
there is no disagreement between the parts of
it.

Page 164

Muqatil on Q. i8:[79] The name of him who used to take each boat

by force was Ibn Qalandl].
(Note: al-Tha'labl reports this name as [Ibn]
Jaladnl)

290

Page 165

Muqatil in Malati

Suyuti

Page 166

Bergstrasser

Ibn Muqaffa'

Page 167

'Umar b. Khattab

gezerah shawah

Page 168

Ibn Anbari

Basran dictum
Ibn Anbari

ANNOTATIONS

Here, according to those who are ignorant in
tafslr, one part contradicts another part. There
is no contradiction, and the two are treated in
tafslr in this way ....

The intention (of Qutrub's work which Suyuti
cites) is to treat the apparent contradictions
between the verses (of the Qur'an). However,
the speech of the Most High is free of that.

because it (^/pas^does not govern the
interpretation of a normative text like the other
(midrashic rules of interpretation), but rather
deals with the extraction of a new regulation
from an existing one.
So he will pay attention in it to the most
truthful of the two parties and to the most
equitable comparison of the two cases.

When there is nothing about it in a Qur'anic
verse or a sunna, (first) consider analogies and
comparisons, and then juxtapose the matters
after that.

inference by analogy

In every analogy, four things are necessary: the
foundation, the derivations, the comparator,
and the ruling.

This is a mere claim for which there is no
evidence except from wahyanA tanztl.
Know that the rejection of analogy in grammar
cannot be proven, because grammar is totally
based on analogy. Those who reject analogy
reject grammar, and no member of the learned
class would reject it.

ANNOTATIONS 291

\ Page 169

Q. 6:82 and have not confounded their belief with

wrong (zuJm)

Q. 31:13 Indeed, association is a mighty wrong {zulm)

Page 171

Muqatil It was revealed as It was prescribed for you (Q.

2:216) — that is, fighting was made obligatory
for you; it was permitted to them after it had
been forbidden to them.

Pages 172—73

Muqatil Muqatil said that on the bridge {jisrj over hell

there are seven bridges (qanatir). At the first
one, the servant is asked about his faith in
God, may He be praised and glorified. If he
answers perfectly and sincerely, he passes on to
the second, at which he is asked about prayer.
If he answers perfectly, he passes on to the
third, at which he is asked about zakat. If he
answers perfectly, he passes on to the fourth, at
which he is asked about fasting. If he answers
perfectly, he passes on to the fifth, at which he
is asked about pilgrimage. If he answers
perfectly, he passes on to the sixth, at which he
is asked about the 'umra. If he answers
perfectly, he passes on to the seventh, at which
he is asked about wrongs. If he has not
committed any wrongs, then he passes into
paradise.

Page 174

Talmudic muttar — "permitted" and asm — "forbidden"
Pauline tupos didaches, "form of teaching"

Page 175

Talmudic formula statutes given to Moses on Sinai

Page 176

Suyutl

translated in the text

2 9 2 ANNOTATIONS

SuyutT translated in the text

SuyutT It is known that its distribution was done

before that although zakat was not yet recited
in revelation, just as ablution was known
before the revelation of the verse. Then the
recitation of the Qur'an was sent down
corroborating it.

Page 178

Suyuti So we bring these reports together with the

multiplicity of (their occasions of) revelation.

Page 180

Ibn Shihab Zuhrl When a sura was revealed in Mecca it was
written in Mecca.

(see A. Rippin, "Al-Zuhrl, naskh al-Qur'an,
and the Problem of Early tafslr Texts," BSOAS
47 [1984]: 22-43 [= A. Rippin, The Qur'an
and Its Interpretative Tradition (Aldershot,
UK, 2001), chap. 16] for the attribution of this
text.)

Ibn 'Abbas on Q. 4:93 Whoever slays a believ er willfully, h is reward is
hell — it was the last of what was revealed, so
nothing can abrogate it.

Page 181

Bukhari Jabra, Mika, and Saraf (mean) "slave" and II

(means) "God."

Page 183

Stetter As I observed; according to how it appeared to

me

Page 184

Q; 2:190 Fight in the way of God with those who fight

with you but aggress not: God does not love
the aggressors.

Q; 9:5 Then, when the sacred months are drawn

away, slay the idolaters wherever you find then,
and take them, and confine them, and lie in

ANNOTATIONS

293

Page 185

Q. 3:200

G\ 5:33

Jassas

Page 187

Goldziher

citing Juwaynl

Page 188

Maimonides

(in Judaeo-Arabic)

0,7:3

Q 59:7

[Jassas]
Abu 'Ubayd

wait for them at every place of ambush. But if
they repent and perform the prayer and pay the
alms, them let them go their way; God is All-
forgiving, All-compassionate.

O you who believe! Be patient and strive in
patience; be steadfast and fear God. Perhaps
you will prosper.

This is the recompense of those who fight
against God and His messenger and hasten
about the earth to do corruption there. They
shall be slaughtered or crucified, or their hands
and feet shall alternately be struck off, or they
shall be banished from the land. That is a
degradation for them in this world and in the
hereafter there will be a mighty chastisement
for them.

(This is) because no ruling is dictated by the
occasion of revelation according to us; rather,
for us, the ruling follows from the general
validity of the scriptural expression.

The zahir is the literal meaning, which is the
(most) probable and preferable {rajih).

These are conclusions that derive from
(tazhuru) from the plain sense of the text.
Follow what has been sent down to you from
your Lord

Whatever the messenger gives you, take, and
whatever he forbids you, give over.
In the obligatory nature of the ruling, it is of a
status with the Qur'an.

The correct text {mahfhz) is, according to us,
(with) the lam.

294 ANNOTATIONS
Page 189

Q. 2:59 Then the evildoers substituted a saying other

than that which had been sent down to them.
ad Q. 2:58 sin: hatt'a; wheat: hitta

Ka'b al-Ahbar The Jews made them into a charm (obliterated

them?) and erased them in order that their
content would not be known.

Q. 5:13 They change {yuharrifhna) the words from

their places.

Page 190

Jassas When we do not find an explicit ruling in the

book or in the sunna or in consensus that will
apply to the incident, because God has said
that in the book there is a clarification of all
things within the realm of religion, it is
established that the way is to use reason and
proof by means of analogy to (achieve) the
ruling about it.

Page 192

Abu 'Ubayd adQ, 5:45

and 2:178 According to our opinion about this verse,

which is in surat al-ma'ida (5), a life for a life,
it is not an abrogator {nasikha) of what is in
surat al-baqara (2) the free for the free and the
slave for the slave. There is no contradiction
here; rather, the two of them agree in being
valid rulings except to the extent that one
considers that (the passage in surat) al-ma'ida
is to be taken as a corroboration {mufassira) of
(the passage in surat) al-baqara. So it would be
explained {taawwala) that (God's) statement a
life for a life applies to the lives of free men
who are equal in what is due to them, but not
to those who are slaves.

Q 5 : 5° Is it the judgment of the jahiliyya that they are

requesting? Yet who is more fair in judgment
than God for a people who are sure in faith?

ANNOTATIONS 295

\ Page 193

Abu TJbayd The Messenger of God said, "Take (it) from

me. God has given them a way, virgin for
virgin, nonvirgin for nonvirgin. Lash, and then
expel, the virgin; lash, and then stone, the
nonvirgin."

Q; 5:42 If they come to you, judge between them or

turn away from them
Q. 5:48 Judge between them according to what God

has sent down.

Abu 'Ubayd The prophet stoned the Jewish man and

woman.

Page 195

Goldziher When four people testify that they saw him

enter into her just as the kohl pencil enters into
the kohl container, then (the penalty is)
stoning.

[This Arabic quote is from Ibn al-Jawzi. It is
printed in Goldziher in Hebrew script (as it
appears here) probably because that was the
typeface available in that journal; it is not
Judaeo- Arabic, of course.]

Page 196

Q. 2:106 And for whatever verse We abrogate or cause

to forget (defer), We bring a better or the like
of it.

ius propter

utilitatem publicam law in the public interest
Hibatallah In the speech of the Arabs (it is) something

removed/ cancelled.

Page 197

Arabic
Tabari

translated in the text

So God lightened (it) for them and abrogated it
by the other verse.

296

ANNOTATIONS

Page 199
Baghdadi

Page 200

Rabbinic formula

antithesis
legislation

Ibn Kammuna

Page 201

Maimonides

Page 203

Abu 'Ubayd

Among them are those who say that, before his
entering into his time of prophethood, the
prophet was ordered to follow the sharfa of
Abraham. He was allowed to hold fast to it in
everything except in what was abrogated of it
in his sharfa after its revelation to him. This is
the sound opinion according to us.

minhag mebattel halakhah, "custom overrides
law"

upheld vs. abolished (bittel) or deleted
minhag, "custom" and derekh eres, "proper
behavior" (literally: the way of the land)
(literally, in Arabic) the spreading of speech;
(Hebrew): bat qoJ, "heavenly speaking voice"

Judaeo-Arabic naskh translated into Hebrew
not as baful (the equivalent of naskh) but as
'ataq (the equivalent of naql)

Chapter on the addition of letters about which
there is disagreement concerning their writing
(in the Qur'an)

Q; 18:79 F° r behind them there was a king who was

seizing every ship in good repair (siliha) by
force.

Muqatil ad Q. 18:79 He takes every shi p — good-repaired (sa/iha), in

good repair (sahlha), sound (sawiyya); by force
(Q. 18:79), just as in God's saying, When He
gave the two of them a salih (Q. 7:190) —
meaning a sound child (sawiy); that is, taken
by force from their people.

Page 204

SuyutI

The goal of the isolated variant is to act as
tafslr of the well-known reading and as an

ANNOTATIONS 297

explanation of its meaning. . . . So, (the
isolated variant) is better and more significant
than tafsir. What is discovered by these words
comes closer to knowledge of sound
interpretation (ta'wil).

Page 206

Suytiti The demonstration of (the Quran's) merit and

honor over all other scriptures is that the other
books that He sent down came with only one
aspect (waj'h).

Page 207

BaydawT He mentions the nights here (Q. 19:10) and

the days in (surat) Al 'Imran (Q. 3:10) to
indicate that the prevention of talking to
people and the devotion to mentioning (God)
and thanking (Him) lasted for three days and
nights.

Page 209

Muqatil The explanation of (the word) wahy provides

five aspects. (1) Among them is that wahy
means what Gabriel brought down from God
to the prophets. That is found in His saying,
We sent revelatio n to you — He means the
Qur'an; just as We sent revelation to Noah and
the prophets after him — then He mentioned
the prophets; and We sent revelation to
Abraham and Ishmael to the end of the verse
(Q. 4:163). He also said, This Qur'an has been
revealed to me so I may warn you thereby (Q.
6:19). There are many similar (verses). (2) The
second aspect is that wahy means "inspiration"
(ilham). That is found in His saying in (surat)
al-ma'ida (Q. 5) And when I inspired
(awhaytu) the disciples — that is, "I inspired
(alhamtu) the disciples"; to believe in Me and
My messenger (Q. 5:111). Similarly He said in
(surat) aI-nahl(Q. 16) Your Lord inspired

ANNOTATIONS

{awha) the bee , — He is saying, "Your Lord
inspired {alhama) the bee; to take a home in
the mountains (Q. 16:68). (3) The third aspect
is that wahy means "writing." That is found in
His saying in {surat) Al 'Imran (Q. 3) to
Zechariah, He wrote {awha} to them — a book;
so they could praise (God) in the morning and
in the evening (Q. 19:11 [surat Maryam]). (4)
The fourth aspect is that wahy means
"command." That is found in His saying in
{surat) Ha-Mim al-sajda (Q. 41), And He
commanded {awha) to each heaven its
command (Q. 41:12). And He said in {surat)
al-An'am (Q. 6), satans among men and jinn
command {yuht) one another — He is saying
that they are commanding {ya'muru) one
another. And He said in surat al-An'am (Q. 6),
The evil ones certainly command {la-yuhuna)
their friends (Q. 6:121) — that is, they order
{ya'muruna) them with whispers and pretense.
(5) The fifth aspect is that wahy means
"speech." That is found in His saying in {surat)
idha zulzilat al-ard (Q. 99), F or that your Lord
will say {awha) to it (Q. 99:5) — that is, He said
{qala) to it.

It may be that they are better than others.
How good is the reward.
God does not forgive associating things with
Him but He does forgive what is other than
that, if He wishes.

The commander does not offer a (single) dinar,
but he does offer great riches ( qintar) to whom
he wishes.

The chapter dealing with single (expressions)
in surat al-baqara (2). O people, serve your
Lord (Q. 2:21), whereas elsewhere in the

ANNOTATIONS

299

Qur'an (He says) Fear your Lord . Also among
them is So bring a sura of a likeness to it (Q.
2:23), whereas elsewhere in the Qur'an (He
says), like it without "of (min). Also among
them is Call your witnesses (Q. 2:23), whereas
elsewhere in the Qur'an (He says), those who
aid you (e.g., Q. 10:37). Ak° among them is
Believe in what I have revealed (Q. 2: 41) in
the first person, whereas elsewhere in the
Qur'an (He says) it with second person. Also
among them is the messengers without cause
(bi-ghayr al-haqq) (Q. 2:61) whereas elsewhere
in the Qur'an (He says), without cause (bi-
ghayr haqq). Also among them is and those
who follow the Jewish way, the Christians and
the Sabians (Q. 2:62), whereas elsewhere in the
Qur'an (He says), and the Sabians comes
before the Christians . Also among them is The
fire shall never touch us except for a few
numbered {ma'duda) days (Q. 2:80), whereas
elsewhere in the Qur'an (He says), a few
numbered {ma 'dudat) . Also among them is
Their penalty shall not be lightened nor shall
they be helped (Q. 2: 86), whereas elsewhere
in the Qur'an (He says), nor shall they
understand (e.g., Q. 2:162). Also among them
is No, but most of them do not believe (Q.
2:100), and there are no other examples of this
in the Qur'an. Also among them is After the
knowledge which has come to you (Q; 2:120)
whereas elsewhere in the Qur'an (He says),
After what has come to you . Also among them
is He will recite to them Your signs and teach
them the book and the wisdom and purify
them (Q. 2:129), whereas elsewhere in the
Qur'an (He says), and purify them and give
them knowledge . Also among them is Do not
fear them but fe ar Me (Q. 2:150) with a ya'
(on the ni of "Me") whereas elsewhere in the
Qur'an it is without the ya'(e.g„ Q. 5:3). Also

3 oo ANNOTATIONS

among them is What has been hallowed of it to
other than God (Q. 2:173), whereas elsewhere
in the Qur'an (He says), to other than God of
it. Also among them is not transgressing, then
no sin shall be on him (Q. 2:173), an d there
are no other examples of this in the Qur'an.
Also among them is Whoever is ill (Q. 2:185)
without adding "among you;" there are no
other examples of it in the Qur'an. Also among
them is Those who believed and those who
emigrated (Q. 2:218) whereas elsewhere in the
Qur'an (He says), and emigrated without
saying "and those who." Also among them is
and will acquit you of your evil deeds (Q.
2:271) whereas elsewhere in the Qur 3 an~(He
says), (acquit) you your evil deeds without
saying "of." Also among them is what we
found {alfayna) o ur fathers doing (Q. 2:170)
whereas elsewhere in the Qur'an (He says), we
found ( wajadna). Also among them is and eat
( wa-Jcula) thereof comfortably (Q. 2:35)
whereas elsewhere (He says), so eat {fa-kula).
Also among them is I_(sic: read "We") said: O
Adam, dwell, you and your spouse (Q. 2:35).
Also among them is No soul is held
responsible (Q. 2:233), ^d there are no other
examples of this in the Qur'an.

Page 214

Q. 2:233 no sou l is held responsible

Q. 2:286, 6:152 He does not hold anyone responsible / We do

not hold responsible

Q. 4:84 You are only held responsible for yourself.

Q; 2:35 We said, "O Adam, dwell, you and your wife,

in the garden, and eat thereof comfortably
wherever you wish. But do not approach this
tree unless you are of the evildoers."

Q. 7:19 O Adam, dwell, you and your wife, in the

garden. So eat of wherever you wish. But do

ANNOTATIONS

not approach this tree unless you are of the
evildoers.

301

Page 217

Ibn 'Abbas
TJmar b. Khattab

translated in the text
translated in the text

Page 218

Ibn 'Abbas

Sufyan al-Thawn

Muqatil
SuyutI

On the authority of 'Ibn 'Abbas concerning the
words of God, In a clear Arabic tongue {lisan
'arabi) [Q. 16:103 etc.]. He said (that this is)
the language of Quraysh. If there had been
other than Arabic in (the Qur'an), (the Arabs)
would not have understood it. God has not
revealed a book without it being in Arabic and
then Gabriel translated it for each prophet into
the language of his people. Therefore God said,
We do not send a messenger unl ess (he comes)
in the language of his people [Q. 14:4]. There
is no language of a people more comprehensive
than the language of the Arabs. There is not in
the Qur'an any language other than Arabic
although that language may coincide with
other languages; however, as for the origin and
category (of the language used), it is Arabic
and nothing is mixed in with it.
(see A. Rippin, "Ibn 'Abbas's al-Lughat M-
Qur'an" BSOAS 44 [1981]: 15-25 [= A.
Rippin, The Qur'an and Its Interpretative
Tradition, chap. 13], for more on this passage.)
Revelation did not come down unless it was in
Arabic. It was then translated to a prophet for
his people.

There are no languages on earth that God has
not included in the Qur'an.
Also the prophet was sent to every community
. . . (Q. 14:4). It is necessary, then, that there
will be in the received book the language of
every community while its grounding is in the
language of the people itself.

302

Page 220

Q. 22:5
Q. 12:82

0; 79 : 3o
Q. 20:69

Page 221

Q. 26:105

Q. 16:66

a 73:18

Q. 55:19-22

Suyuti

Suyuti

Seven canonical
readings

Page 222

Q. 20:63
Suyuti (b)
Suyuti (c)
Suyuti (e)
Ibn Qutayba
Ibn Qutayba

Ibn Qutayba

Ibn Qutayba

ANNOTATIONS

We bring you forth as an infant (as infants).
And ask (the people of) the town in which we
were and (those in) the caravan with which we
proceeded.

the earth, after (with) that He spread it out
What they have done (their deed) is the trick
of a sorcerer.

Noah's people accused (kadhdhabat I

kadhdhaba) the messengers of lying.

In the cattle there is a lesson for you: We give

you to drink of what is in their bellies

(bufunihi / butuniha).

Heaven (the ceiling) is rent thereby.

He has loosed the two seas to meet . . . pearls

large and small come forth from both of them

(one of them).

('A'isha) This is the work of the scribes who
erred in writing the book; (TJthman) Do not
change it; the Arabs will change it.
Its wording differs from its writing; multiple
aspects in the readings

The best choice among the seven readings

These are two sorcerers {in hadhani la-sahirani)

innahu hadhani la-sahirani

innahu hadhani lahuma sahirini

innaha dhani la-sahirani

(Ubayy) in dhani illa-sahirani

(Ibn Mas'ud) an hadhani sahirini or in

hadhani la-sahirani or in hadhani sahirini

('Asim) writes inna hadhani, reads inna

hadhayni

The difference between the reading and the
book according to the statement of TJthman,
may God have mercy on him: "I see errors in

I

ANNOTATIONS

303

Abu TJbayda

Page 223

Hebrew
Arabic
al tiqra

tiqqunei soferim
miqra soferim

Page 224

Ibn Munayyir
Q. 19:9

Hebrew
Zamakhsharl

Q. 17:190

it, and the Arabs will establish it according to
their dialects." So he established it according to
his dialect and left the writing as it was.
(It is) in hadhayni la-sahirani according to the
(pronounced) wording {Iafz) but it is written
hadhani, just as (letters) are added to and taken
away from the book (elsewhere). But the
(pronounced) wording remains preferable.

ketib : qere (written : recited)

written : pronounced wording; book : reading

do not read

emendations of the scribes
recitation of the scribes

translated in the text

He said / they said, "Thus {kadhalika) Your
Lord says."

koh amar Yahweh, "So said the Lord"
In kadhalika, the kaf\\a.s an "a" vowel; that is,
(it means) "the command is thus," that being a
confirmation of (what has been said). Then He
continued (after the disjunction) with Your
Lord said. Or, {kadhalika) may be understood
(because of the "a" vowel) as the object of (He)
said .

God is exalted above what they associate (with
Him).

Page 225

Q. 106:1
Proverbs 26:23
Brockelmann

The caravans of Quraysh

"like glaze" vs. "unrefined silver"

the inclination of Semitic languages to form

simple sentences, always dividing complicated

notions into simple formulations

3 o 4 ANNOTATIONS
Page 227

Sharif Murtada Know that among the customs of the Arabs are

concision, brevity, ellipsis, searching for the
shortness of expression and repudiation of
excess, being satisfied with brevity rather than
long-windedness. . . .When you contemplate
the kinds of figurative expressions that linguists
use in their poems and prose, you will find all
of them explained by ellipsis and brevity, and
likewise, in God's saying and your Lord comes
(Q. 89:22) and ask the town (Q. 12:82),
ellipsis in them is obvious.

Page 228

Ibn Qutayba As for those who try to discredit the Qur'an for

its figural speech {majaz), they claim that (the
Qur'an) tells lies because a wall cannot wish
(Q. 18:77) and a town cannot be asked (Q.
12:82). This just shows the depths of their
ignorance and is suggestive of their lack of
understanding and limited intelligence. If
majaz is considered lying and all actions
attributed to nonanimate beings are considered
false, then most of our speech would be corrupt
because we speak of "plants sprouting," "trees
growing tall," "dates ripening," "mountains
arising," and "prices going down."

Page 229

Q; 55:31 And we will attend to you.

Ibn Qutayba We will begin with the chapter on metaphor

because that is the most commonly occurring

(type of) majaz.

Q. 68:42 Upon the day on which the shank shall be laid

bare.

Ibn Qutayba That is, on account of the strength of the

matter, he bared his shank; so "shank" stands
(/a-'stu c irat) in place of "strength."

Ibn Qutayba
adQ. 17:23

Ibn Qutayba/Jurjanl

Page 230

Ibn Qutayba

Page 231

Ibn Qutayba
Ibn Qutayba

Page 232

Baqillanl/

Imru' al-Qays

Ibn al-Mu'tazz

Q. 21:22

Page 233

Subki

ANNOTATIONS

Th.

305

e root of this is your blowing any dirt or
ashes or anything similar which has fallen on
you.

that one omits the thing that possesses (i.e., the
"people" who belong to the town who must do
the asking in Q. 12:82) and puts the thing
possessed ("the town") in its place and makes it
the subject of (the statement).

Inversion is when something is described by a
characteristic that is its opposite.

He alluded ( warra) to their remembering.
Where the meaning differs from the obvious
wording

What the south wind and the north wind have

woven over it (li-ma nasajatha ought to have

read li-ma nasajaha).

I was extravagant in secrecy;

which was on my part flattery.

I kept secret my love for you to a point where;

I was keeping secret my secrecy.

So that no course was open to me;

Other than to mention it with my tongue.

translated in the text

With you is the contrary of the contrary of
that;

in which is the contrary of the contrary of
beautiful.

His beauties are the essence of all beauty;
And the hearts of men are magnets.

3 o6 ANNOTATIONS
Page 234

Q. 28:72 Of His mercy He has appointed for you night

and day for you to repose in and to seek His
bounty.

Q; 30:23 And of His signs are your slumbering by night

and day and seeking His bounty.

Q. 2:1 1 1 And they say, "None shall enter paradise unless

they are Jews or Christians."

Q. 2:185 Those of you who are present at the month,

fast it; and those who are sick or on a journey

Q. 13:12 It is He who shows you the lightning for fear

and hope.

(Ibn Hayyus) How can I forget, when you are a dune, a

branch,

and a sun, in glance, stature, and figure?

Page 235

Q. 7:89 It is not for us to return into it unless God, our

Lord, wills (it).

Page 236

Q. 42:1 1 There is nothing like unto Him (ka-mithlihi).

JurjanI The least that the first group — those who deny

majaz — should come to understand is that
revelation {tanzil) did not change the specific
laws governing the operation of language and it
did not give different meanings to words. . . .
Thus, revelation did not attempt to change the
habits (of speech) of its people, nor did it
attempt to move them to different styles and
modes of expression. Revelation certainly did
not prohibit them from recognizing that which
they already know of (their language),
including simile, parable, ellipsis, and allusion.

Page 238

Sister

Nature: heaven and the stars, storms, colors,
scenery, animal world, plants; Man and his life:

ANNOTATIONS

307

Sabbagh
Page 239

0:62=5

Hebrew (Geiger)
Q: 30:58

Q, 17:89
0; 24:34

Ezekiel 14:8
Page 240

Q; 2533

Page 241

Zamakhsharl

Zamakhshari

Page 242

Suyuti

parts of the body, the family, messengers,
society, farming, arts, and crafts
Nature: man, parts of the human body,
function, and actions of the body; social life

The likeness of those who carried the Torah

and then did not carry it is as the likeness of

the donkey who carries books.

a donkey carrying books

Indeed, We have struck {darabna) for the

people in this Qur'an every manner of

similitude (mathal).

We have turned about {sarrafna) for the people
in this Qur'an every manner of similitude
{mathal).

We have sent down to you signs (ayat) making

all clear, and a similitude {mathal) of those

who passed away before you, and an

admonition for the god-fearing.

I will make him a sign (of) and a proverb

{meshalim).

They do not come to you with a similitude
{mathal) except We bring you the truth {haqq)
and better in exposition.

This is a figure which is named tamthll,
although it is also possible that it is takhyll.
It is tamthil 'and takhyil.

Amthal provide an aid to comprehension
through personification because parables are
established in the mind by making use of sense
data.

308 ANNOTATIONS

Q. 29:43 And those similitudes, We strike them for the

people, but no one understands them other
than those who know.

Sahl al-Tustari All verses of the Qur'an have four senses

(ma'an): the literal and the symbolic, the
prescriptive and the spiritual. The literal is
(what is) recited and the symbolic is (what is)
understood; the prescriptive is the permitted
and the forbidden, while the spiritual is the
opening of the heart to what is intended for it
as insight from God.

Page 243

Suyuti Texts have their literal meanings and along

with that (they have) concealed signs pointing
to secret implications (or, if read, haqa'iq, as in
footnote, then "truths" or "realities" or, as
suggested on the next page of the text "arcane
meanings").

Page 246

Qphelet 8:1 Who is like the wise man and who knows the

interpretation of a matter?

GLOSSARY

id hominem

aetiological

anacolutha
antonomasia,
antonomastic

apparatus criticus
apodictic

aphorism, aphoristic
apostrophe

argumentum
ad hominem

argumentum e silencio
autochthonic

rejection of an argument on the basis of an
irrelevant fact about the author of the
argument

providing an explanation of origins in the form
of a narrative

sentences lacking grammatical sequence

substitution of an epithet etc. for a proper
name

textual critical apparatus

providing a very strong imperative, as in a

divine saying which has the power to transform

reality, illustrated by the biblical Ten

Commandments.

short, instructive saying

figure of speech addressing an absent or dead
person, thing, or idea as if it were alive and
present

irrelevant attack upon a person to deflect an

argument from fact and reason

argument from silence

of one who originates from where he/she is

found

benedictio

Bildungserlebnis

Bildungsprinzip

Botenformel

Botenspruch

caique

ceremonial prayer invoking divine protection
foundational or educational event or experience
instrument of education
messenger formula
utterance (of the messenger)

the adoption by one language of a word whose
components are literal translations of the
components of a corresponding word in

3io

GLOSSARY

canon, canonicity

casus obliquus
casus rectus
chiasmus

another language (as in the English "hot dog"
being rendered in French as "chien chaud")
writing constituting a collection of texts
accepted by religious leaders and communities
as authoritative scripture, suggesting a process
of selection that deemed some texts more
worthy of preservation than others; belonging
to that canon.

in grammar, the accusative or genitive case
in grammar, the nominative case
a literary structure in which elements are
repeated in reverse, giving the pattern ABBA
constructio ad sensum structured according to the sense (and not

grammar)
contra naturam against custom

corroboratio formula of legitimation

deictic demonstrative (grammatical); word specifying

identity or spatial or temporal location from
the perspective of a speaker or hearer

Deutungsbedurftigkeit the need for interpretation

diatheke covenant

diglossia existence of two official languages in a society

disputatio fori debate in the court of justice/forum

e silencio
enjambment

epexegetic

epistolaris sermo
epitheta ornantia
erlebte Rede
Evangelion
evangelium
exemplum, exempla

exotica

from silence

continuation of a complete idea or grammatical
structure from one line of a poem to the next
providing additional explanation or explanatory
material

letterlike sermon

decorative phrase

narrated monologue

literary genre of "gospel"

gospel (as a narrative form of biography)

anecdote or short narrative used for a moral or

in an argument

foreign

fictio personae creation of a person treated as being present

GLOSSARY

311

figura etymologica
fissiparous

Gemeindebildung

haggadic

halakhic

hapax legomenon
(pi: legomena)

Heilsgeschichte

hendiadys

historia

homoioteleuton

hortatory
hyperbaton

inconcinnity

inter alia
inter alios
invocatio

juncture

figure of speech based upon a word etymology
fruitful, generative

community formation or education

the part of traditional Jewish literature
consisting of elaborations on the biblical
narratives or tales from the lives of the Rabbis
and, thus, in reference to any traditional
homiletic, sermonic interpretation of scripture
(other common spelling: Aggadic)
Jewish legal literature and, thus, in reference to
any traditional legal interpretation of scripture
(other spelling: Halachic)

an isolated use of a word, there being no other
words related to that word present in the text
salvation history

use of two words plus "and" for a single idea
historical narrative

literally, "same ending"; common scribal error
occurring when two words/phrases/lines end
with the same sequence of letters and the
scribe, having finished copying the first, skips
to the second, omitting all intervening words
giving exhortation or advice; encouraging;
inversion of normal word order

needing concinnity or congruousness;
unsuitableness
among other things
among other persons

appeal for inspiration made to a muse or deity,
usually at the beginning of a work

(in linguistics) a pause in speech or a feature of
pronunciation that functions as a pause

kerygma

proclamation; used to refer to the content of
early Christian preaching

312

koine

Kunstprosa

langue courante
lapsus calami
laudatio
lex talionis
lexis

lingua sacra

litteratim

loci

locus classicus

locus probans/

loci probantes
logia

GLOSSARY

the "common dialect" of Greek that flourished
under the Roman Empire, bringing together
features of a number of dialects and used by
speakers of different languages (as found in the
New Testament)

elevated discourse that is between poetry and
prose

language in use
an error in writing
commendation; eulogy
law of retaliation
lexical meaning
sacred, liturgical language
letter for letter
passages

passage which is used as a standard example
for the elucidation of a word or subject

places of use, proof texts

oracles, divine utterances, specifically sayings of
Jesus that lie behind the Gospels

masora, masoretic,
masoretes

me

tath

esis

mimesis, mimetic

mise en scene
Mitteilungsbedurfnis

the masora (or Massorah) is the text of the
Hebrew Bible as officially revised by the
Masoretes from the sixth to the tenth centuries
CE, with critical notes and commentary and,
thus, the collection of these notes and
commentaries

transposition of sounds or syllables within a
word

the imitative representation of nature and
human behavior in art and literature, especially
the representation of another person's words in
a speech

setting; environment of an event
normal communication

GLOSSARY

3i3

I mnemonic

mythopoeic
Nachdichtung

narratio
ne varietur
nomen regens

nomen rectum

nomina propria
non sequitur

onomatopoeic
ornatus

paideutic
parabolic

paraenesis, paraenetic
paraphrastic
paronomasia,
paronomastic

per antiphrasin

pericope

pleonasm, pleonastic
point d'appui

preciosity

word, abbreviation, rhyme, or similar verbal
device that is learned or created in order to
remember something

artistic reimagining of mythological narratives

reformulation of an earlier text in a manner
that the author considers more fitting for the
new audience; a term used for early twentieth-
century German translations of poetry,
especially connected to the "rendering" by Karl
Kraus of Shakespeare into German,
exposition or explanation of a topic
standard, not changing

the modified or defined noun in a grammatical

possessive construction

the modifying or defining noun in a

grammatical possessive construction

proper nouns

a reply that has no relevance to what preceded
it or a conclusion that does not follow from its
premises

formed in imitation of a natural sound
(rhetorical) flourishes

educative

expressed in a parable
exhortation, exhortative
paraphrased in its meaning

pun; punning; often, recurrence of the same

word or word stem within a sentence

use of words in a sense opposite to the proper

meaning

a selection used in worship or instruction, as in

a Bible section used in worship

words that are superfluous

literally "the point of support"; a position from

which one may proceed

overrefinement in language and manners

3M

GLOSSARY

primum comparationis the first element of a comparison

proleps

is

Prophetenkultus

propinquity

prosopopoeia

ratio

schemata

scriptio defectiva

secundum

comparationis
semasiological
sermo plebeius
sermo urbanus
Sitz im Leben

Stichworte

Straflegenden

sub specie aeternitatis

sui

generis

Siindenregister
synaesthetic

anticipating and answering objections in
advance

cult of the prophet
being close together

representing an abstract quality or idea as a
person

(in law) underlying reasoning or principle

the structures or units into which knowledge is
organized

defective writing due to the nature of a written
script

the second element of a comparison
semantic

speech of the common person
cultured speech

literally, situation in life or setting;
fundamental concept of biblical Form Criticism
that looks at the structure and intention of the
texts (with an emphasis on the oral
transmission and productions of texts) to come
to understand the sociological setting of the
literary genres employed,
symbols

punishment stories, retribution pericopes
literally "under the aspect of eternity"; thus, an
expression describing what is universally true
without any reference to temporal reality
literally "of its own kind"; thus, whatever is
totally unique or distinctive
literally, "list of sins"; compilation of errors
involving more than one sense

terminus ad quem
terminus technicus

final or latest limiting point in time
technical terminology

GLOSSARY

3i5

tertium comparationis

theodicy
theologoumena

tmesis

topos (pi. topoi)
triptosy

typology, typological

Ubi sunt

Urerlebnis
Urgemeinde

Urkundenhypothese

Urtext

usus loquendi

varia lectio,

variae lectiones
vaticinatio
vaticinatio ex eventu
versus rapportati

virtus

Volkerwanderung

third term in a comparison; the quality that
two things which are being compared have in
common

argument and justification for belief in God
theological interpretations that are subject to
debate but are accepted as doctrine because
they are closely aligned to a defined doctrine
about God

separation of parts of words by an intervening
word

conventionalized expression or passage in a
text used as a basis for composition of a
subsequent text

three possible grammatical endings of a noun
form of interpretation wherein a person, event,
or institution is viewed as foreshadowing a later
one

a poetical motif emphasizing the transitory

nature of youth, life, and beauty

original event or experience

the foundational Christian community in

Jerusalem

documentary hypothesis (in biblical studies
regarding the composition of the Torah)
original text
usage in speaking

variant reading(s) of the text

prophecy, prediction

prophecy after the event

a literary figure in poetry where the

components correspond; for example:

Aire, Water, Earth

By Fowl, Fish, Beast

Was flown, was swum, was walkt
(see Wansbrough, BSOAS31 (1968): 473)
strength, power, virtue
nomadism

3 i6 GLOSSARY

Vorbild background

Vorlage model or prototype

Wahrheit (historical) truth

Wirklichkeit (historical) fact

zeugma use of a word to govern two or more words

though appropriate to only one

Quranic Studied deserves serious atten-
tion, as a stimulating work ol scholarship.
Its allusions to biblical and Arabian under-
pinnings have captured main' people's at-
tention and led to numerous exchanges and
debates among scholars and others, espe-
cially regarding Wansbrough's claim that
the Qur'an was not written down until the
third-century />///•/ (ninth century CI.),
countering traditional Muslim claims that
it originated in the time ol Muhammad and
was written down shortly thereafter. In
response, some decried the publication of
Quranic Studied, seeing it as a major imped-
iment to fostering a trust of nonsectarian
scholarship among Muslims. Now readers
can judge lor themselves.

Although Quranic Studied was origi-
nally intended for fellow scholars, the In-
ternet has considerably widened its acces-
sibility. Used as a source of authority and
critical opinion on polemical sites by both
Muslims and Christians, Wansbrough's
work has gained a significant profile among
professionals and nonprofessionals alike.
Nonetheless, it appears that many who cite
Quranic Studied have not carefully read it.
To counter such ideological and nonschol-
arly treatments of Wansbrough's ideas,
noted Islamic scholar Andrew Rippin has
enhanced the work with a valuable tore-
word, helpful text annotations, and a much-
needed glossary to increase the utility of
this seminal work for the many avid read-
ers w ho desire to know more about Islam.

y image 20QQ Tom C«mpa§Mmii WAX VISUAL
./ticket tk\>ian /'(/ Grace ill. L\mli-Zit.<l>er<ier

John N\ anshrough contributed 10 various as|M-cts <>l the history anil culture <»! i In- Mi. I
die Host and the Mediterranean, Inii lie was most important in stimulating new approaches
to the study ol early Islam. Starting Trout the view that Islam's own accounts ol its origins
are religiously inspired interpretations ol history rather than true records ol events, Wans-
hrough developed new and influential theories ahout the ways in which the emergence of
Islam should he understood and studied.

"I le was Ih'sI known lor his work on the Koran. In his (Jitranif ShuHe,: using his pro-
lound knowledge ol classical and modern literary and historical theory, he applied to the
holy IxKtk ol Islam ideas and approaches w hich scholars had dcvclo|>ed in the study ol die
Mible and early Christianity. Mis literary- analysis id the Koran and the commentaries on
it led him to v iews very dilTerenl from those held hy traditional Muslims and by the ma-
jority ol non-Muslim academics.

M . . . | YYanshrough | questioned the accepted \ iew- that the Koran consists ol passages
associated with (or, in the traditional understanding, divine revelations mode to) Muham-
mad in Mecca and Medina in the early decades ol the seventh 08QtOiy« that it had been com-
mitted to writing' by alniul (>f>0, and that it was the most important element in Islam from
the lime of Muhammad onwards.

" These and the other standard ways of approaching the Koran, he argued, resulted
from too w illing an acceptance of Islam's own tradition — primarily the Itody of traditional
Muslim commentary on the Koran.

". . . Starting from the liasis that there is very little Islamic literature securely datable
to before about N00, Wanshrnugh saw Islam as evolving gradually from sectarian forms of
Judaism over a |>eriod ol I AO years or so following the* Arab conquests in the middle ol
the seventh century.

"lie understood the history of that formative |>eriod, including the image of Muham-
mad and accounts ol the lormation of the Koran, as a back projection ol views that were
formed as the culture and religion of Islam emerged in an atmosphere of intense polemic
between different groups of monot heists. . . .

"Mis views and approaches remain controversial, both in academic circles and, lor
more obvious reasons, aiming Muslims who know of them, lint they were not expressed out
ol any hostility to Islam. On the contrary, Wanshrnugh insisted that, together with Juda-
ism and Christianity. Islam is a valid expression of the monotheistic tradition of religion and
that it must Ik- treated with the same scholarly seriousness as its sisters."

— / /•. '///««■.« ( I i. ii. Inn i

Prometheus Books

59 John Giefln Drive • Amherst. New York 14228-2 W
\v>w,v\.iiri.\)Metlvui»biwils. 4 dm